Authors: Siddhant Dutta, Hemant Singh, Kalpita Shankhdhar, Sridhar Iyer
Abstract: Consuming poisonous mushrooms can have severe health consequences, even resulting in fatality and accurately distinguishing edible from toxic mushroom varieties remains a significant challenge in ensuring food safety. So, it's crucial to distinguish between edible and poisonous mushrooms within the existing species. This is essential due to the significant demand for mushrooms in people's daily meals and their potential contributions to medical science. This work presents a novel Quantum Vision Transformer architecture that leverages quantum computing to enhance mushroom classification performance. By implementing specialized quantum self-attention mechanisms using Variational Quantum Circuits, the proposed architecture achieved 92.33% and 99.24% accuracy based on their category and their edibility respectively. This demonstrates the success of the proposed architecture in reducing false negatives for toxic mushrooms, thus ensuring food safety. Our research highlights the potential of QMViT for improving mushroom classification as a whole.
Authors: Thao Le, Tim Miller, Ruihan Zhang, Liz Sonenberg, Ronal Singh
Abstract: This paper presents Visual Evaluative AI, a decision aid that provides positive and negative evidence from image data for a given hypothesis. This tool finds high-level human concepts in an image and generates the Weight of Evidence (WoE) for each hypothesis in the decision-making process. We apply and evaluate this tool in the skin cancer domain by building a web-based application that allows users to upload a dermatoscopic image, select a hypothesis and analyse their decisions by evaluating the provided evidence. Further, we demonstrate the effectiveness of Visual Evaluative AI on different concept-based explanation approaches.
Authors: Jiajia Li, Kyle Lammers, Xunyuan Yin, Xiang Yin, Long He, Renfu Lu, Zhaojian Li
Abstract: Fruit harvesting poses a significant labor and financial burden for the industry, highlighting the critical need for advancements in robotic harvesting solutions. Machine vision-based fruit detection has been recognized as a crucial component for robust identification of fruits to guide robotic manipulation. Despite considerable progress in leveraging deep learning and machine learning techniques for fruit detection, a common shortfall is the inability to swiftly extend the developed models across different orchards and/or various fruit species. Additionally, the limited availability of pertinent data further compounds these challenges. In this work, we introduce MetaFruit, the largest publicly available multi-class fruit dataset, comprising 4,248 images and 248,015 manually labeled instances across diverse U.S. orchards. Furthermore, this study proposes an innovative open-set fruit detection system leveraging advanced Vision Foundation Models (VFMs) for fruit detection that can adeptly identify a wide array of fruit types under varying orchard conditions. This system not only demonstrates remarkable adaptability in learning from minimal data through few-shot learning but also shows the ability to interpret human instructions for subtle detection tasks. The performance of the developed foundation model is comprehensively evaluated using several metrics, which outperforms the existing state-of-the-art algorithms in both our MetaFruit dataset and other open-sourced fruit datasets, thereby setting a new benchmark in the field of agricultural technology and robotic harvesting. The MetaFruit dataset and detection framework are open-sourced to foster future research in vision-based fruit harvesting, marking a significant stride toward addressing the urgent needs of the agricultural sector.
Authors: Oresti Banos, Zhoe Comas-Gonz\'alez, Javier Medina, Aurora Polo-Rodr\'iguez, David Gil, Jes\'us Peral, Sandra Amador, Claudia Villalonga
Abstract: Background: Human Emotion Recognition (HER) has been a popular field of study in the past years. Despite the great progresses made so far, relatively little attention has been paid to the use of HER in autism. People with autism are known to face problems with daily social communication and the prototypical interpretation of emotional responses, which are most frequently exerted via facial expressions. This poses significant practical challenges to the application of regular HER systems, which are normally developed for and by neurotypical people. Objective: This study reviews the literature on the use of HER systems in autism, particularly with respect to sensing technologies and machine learning methods, as to identify existing barriers and possible future directions. Methods: We conducted a systematic review of articles published between January 2011 and June 2023 according to the 2020 PRISMA guidelines. Manuscripts were identified through searching Web of Science and Scopus databases. Manuscripts were included when related to emotion recognition, used sensors and machine learning techniques, and involved children with autism, young, or adults. Results: The search yielded 346 articles. A total of 65 publications met the eligibility criteria and were included in the review. Conclusions: Studies predominantly used facial expression techniques as the emotion recognition method. Consequently, video cameras were the most widely used devices across studies, although a growing trend in the use of physiological sensors was observed lately. Happiness, sadness, anger, fear, disgust, and surprise were most frequently addressed. Classical supervised machine learning techniques were primarily used at the expense of unsupervised approaches or more recent deep learning models.
Authors: Isha Garg, Christian Koguchi, Eshan Verma, Daniel Ulbricht
Abstract: In this work, we study how well the learned weights of a neural network utilize the space available to them. This notion is related to capacity, but additionally incorporates the interaction of the network architecture with the dataset. Most learned weights appear to be full rank, and are therefore not amenable to low rank decomposition. This deceptively implies that the weights are utilizing the entire space available to them. We propose a simple data-driven transformation that projects the weights onto the subspace where the data and the weight interact. This preserves the functional mapping of the layer and reveals its low rank structure. In our findings, we conclude that most models utilize a fraction of the available space. For instance, for ViTB-16 and ViTL-16 trained on ImageNet, the mean layer utilization is 35% and 20% respectively. Our transformation results in reducing the parameters to 50% and 25% respectively, while resulting in less than 0.2% accuracy drop after fine-tuning. We also show that self-supervised pre-training drives this utilization up to 70%, justifying its suitability for downstream tasks.
Authors: Kambiz Azarian, Debasmit Das, Qiqi Hou, Fatih Porikli
Abstract: We introduce segmentation-free guidance, a novel method designed for text-to-image diffusion models like Stable Diffusion. Our method does not require retraining of the diffusion model. At no additional compute cost, it uses the diffusion model itself as an implied segmentation network, hence named segmentation-free guidance, to dynamically adjust the negative prompt for each patch of the generated image, based on the patch's relevance to concepts in the prompt. We evaluate segmentation-free guidance both objectively, using FID, CLIP, IS, and PickScore, and subjectively, through human evaluators. For the subjective evaluation, we also propose a methodology for subsampling the prompts in a dataset like MS COCO-30K to keep the number of human evaluations manageable while ensuring that the selected subset is both representative in terms of content and fair in terms of model performance. The results demonstrate the superiority of our segmentation-free guidance to the widely used classifier-free method. Human evaluators preferred segmentation-free guidance over classifier-free 60% to 19%, with 18% of occasions showing a strong preference. Additionally, PickScore win-rate, a recently proposed metric mimicking human preference, also indicates a preference for our method over classifier-free.
Authors: Sree Rama Vamsidhar S (Indian Institute of Technology), Rama Krishna Gorthi (Indian Institute of Technology)
Abstract: In this paper, we propose to reformulate the blind image deblurring task to directly learn an inverse of the degradation model using a deep linear network. We introduce Deep Identity Learning (DIL), a novel learning strategy that includes a dedicated regularization term based on the properties of linear systems, to exploit the identity relation between the degradation and inverse degradation models. The salient aspect of our proposed framework is it neither relies on a deblurring dataset nor a single input blurred image (like Polyblur, a self-supervised method). Since it is purely image-data-independent, we term our model as Null-Shot deblurring Using Deep Identity Learning (NSD-DIL). We also provide an explicit representation of the learned deep linear network in a matrix form, called Deep Restoration Kernel (DRK) for deblurring task. The proposed framework detours the typical degradation kernel estimation step involved in most of the existing blind deblurring solutions by the proposition of our Random Kernel Gallery (RKG) dataset. In this work, we focus on the restoration of mild blur images, generated by small out-of-focus, lens blur, or slight camera motion, which often occurs in real images. Our experiments show that the proposed method outperforms both traditional and deep learning based deblurring methods, with at least an order of 100 lesser computational resources. The proposed NSD-DIL method can be effortlessly extended to the Image Super-Resolution (ISR) task as well to restore the low-resolution images with fine details. The NSD-DIL model and its kernel form representation (DRK) are lightweight yet robust and restore the mild blur input in a fraction of a second. Hence, more suitable for wide real-time applications.
Authors: Younggun Kim (Department of Civil Engineering, University of Central Florida, Florida, USA), Beomsik Cho (Department of AI Mobility Engineering, Suwon, South Korea), Seonghoon Ryoo (Department of AI Mobility Engineering, Suwon, South Korea), Soomok Lee (Department of AI Mobility Engineering, Suwon, South Korea)
Abstract: Adapting deep learning networks for point cloud data recognition in self-driving vehicles faces challenges due to the variability in datasets and sensor technologies, emphasizing the need for adaptive techniques to maintain accuracy across different conditions. In this paper, we introduce the 3D Adaptive Structural Convolution Network (3D-ASCN), a cutting-edge framework for 3D point cloud recognition. It combines 3D convolution kernels, a structural tree structure, and adaptive neighborhood sampling for effective geometric feature extraction. This method obtains domain-invariant features and demonstrates robust, adaptable performance on a variety of point cloud datasets, ensuring compatibility across diverse sensor configurations without the need for parameter adjustments. This highlights its potential to significantly enhance the reliability and efficiency of self-driving vehicle technology.
Authors: Zhaorun Chen, Yichao Du, Zichen Wen, Yiyang Zhou, Chenhang Cui, Zhenzhen Weng, Haoqin Tu, Chaoqi Wang, Zhengwei Tong, Qinglan Huang, Canyu Chen, Qinghao Ye, Zhihong Zhu, Yuqing Zhang, Jiawei Zhou, Zhuokai Zhao, Rafael Rafailov, Chelsea Finn, Huaxiu Yao
Abstract: While text-to-image models like DALLE-3 and Stable Diffusion are rapidly proliferating, they often encounter challenges such as hallucination, bias, and the production of unsafe, low-quality output. To effectively address these issues, it is crucial to align these models with desired behaviors based on feedback from a multimodal judge. Despite their significance, current multimodal judges frequently undergo inadequate evaluation of their capabilities and limitations, potentially leading to misalignment and unsafe fine-tuning outcomes. To address this issue, we introduce MJ-Bench, a novel benchmark which incorporates a comprehensive preference dataset to evaluate multimodal judges in providing feedback for image generation models across four key perspectives: alignment, safety, image quality, and bias. Specifically, we evaluate a large variety of multimodal judges including smaller-sized CLIP-based scoring models, open-source VLMs (e.g. LLaVA family), and close-source VLMs (e.g. GPT-4o, Claude 3) on each decomposed subcategory of our preference dataset. Experiments reveal that close-source VLMs generally provide better feedback, with GPT-4o outperforming other judges in average. Compared with open-source VLMs, smaller-sized scoring models can provide better feedback regarding text-image alignment and image quality, while VLMs provide more accurate feedback regarding safety and generation bias due to their stronger reasoning capabilities. Further studies in feedback scale reveal that VLM judges can generally provide more accurate and stable feedback in natural language (Likert-scale) than numerical scales. Notably, human evaluations on end-to-end fine-tuned models using separate feedback from these multimodal judges provide similar conclusions, further confirming the effectiveness of MJ-Bench. All data, code, models are available at https://huggingface.co/MJ-Bench.
Authors: Juheon Lee, Xiaohao Cai, Carola-Bibian Sch\"onlieb, Simon Masnou
Abstract: Point clouds are popular 3D representations for real-life objects (such as in LiDAR and Kinect) due to their detailed and compact representation of surface-based geometry. Recent approaches characterise the geometry of point clouds by bringing deep learning based techniques together with geometric fidelity metrics such as optimal transportation costs (e.g., Chamfer and Wasserstein metrics). In this paper, we propose a new surface geometry characterisation within this realm, namely a neural varifold representation of point clouds. Here the surface is represented as a measure/distribution over both point positions and tangent spaces of point clouds. The varifold representation quantifies not only the surface geometry of point clouds through the manifold-based discrimination, but also subtle geometric consistencies on the surface due to the combined product space. This study proposes neural varifold algorithms to compute the varifold norm between two point clouds using neural networks on point clouds and their neural tangent kernel representations. The proposed neural varifold is evaluated on three different sought-after tasks -- shape matching, few-shot shape classification and shape reconstruction. Detailed evaluation and comparison to the state-of-the-art methods demonstrate that the proposed versatile neural varifold is superior in shape matching and few-shot shape classification, and is competitive for shape reconstruction.
Authors: Kenneth D. Forbus, Kezhen Chen, Wangcheng Xu, Madeline Usher
Abstract: One of the purposes of perception is to bridge between sensors and conceptual understanding. Marr's Primal Sketch combined initial edge-finding with multiple downstream processes to capture aspects of visual perception such as grouping and stereopsis. Given the progress made in multiple areas of AI since then, we have developed a new framework inspired by Marr's work, the Hybrid Primal Sketch, which combines computer vision components into an ensemble to produce sketch-like entities which are then further processed by CogSketch, our model of high-level human vision, to produce both more detailed shape representations and scene representations which can be used for data-efficient learning via analogical generalization. This paper describes our theoretical framework, summarizes several previous experiments, and outlines a new experiment in progress on diagram understanding.
Authors: Emma Andrews, Prabhat Mishra
Abstract: Image classification is an essential part of computer vision which assigns a given input image to a specific category based on the similarity evaluation within given criteria. While promising classifiers can be obtained through deep learning models, these approaches lack explainability, where the classification results are hard to interpret in a human-understandable way. In this paper, we present an explainable metric learning framework, which constructs hierarchical levels of semantic segments of an image for better interpretability. The key methodology involves a bottom-up learning strategy, starting by training the local metric learning model for the individual segments and then combining segments to compose comprehensive metrics in a tree. Specifically, our approach enables a more human-understandable similarity measurement between two images based on the semantic segments within it, which can be utilized to generate new samples to reduce bias in a training dataset. Extensive experimental evaluation demonstrates that the proposed approach can drastically improve model accuracy compared with state-of-the-art methods.
Authors: Zaid A. El-Shair, Abdalmalek Abu-raddaha, Aaron Cofield, Hisham Alawneh, Mohamed Aladem, Yazan Hamzeh, Samir A. Rawashdeh
Abstract: Robust perception is critical for autonomous driving, especially under adverse weather and lighting conditions that commonly occur in real-world environments. In this paper, we introduce the Stereo Image Dataset (SID), a large-scale stereo-image dataset that captures a wide spectrum of challenging real-world environmental scenarios. Recorded at a rate of 20 Hz using a ZED stereo camera mounted on a vehicle, SID consists of 27 sequences totaling over 178k stereo image pairs that showcase conditions from clear skies to heavy snow, captured during the day, dusk, and night. The dataset includes detailed sequence-level annotations for weather conditions, time of day, location, and road conditions, along with instances of camera lens soiling, offering a realistic representation of the challenges in autonomous navigation. Our work aims to address a notable gap in research for autonomous driving systems by presenting high-fidelity stereo images essential for the development and testing of advanced perception algorithms. These algorithms support consistent and reliable operation across variable weather and lighting conditions, even when handling challenging situations like lens soiling. SID is publicly available at: https://doi.org/10.7302/esz6-nv83.
Authors: Haolin Pan, Yong Guo, Mianjie Yu, Jian Chen
Abstract: Real-world data often follows a long-tailed distribution, where a few head classes occupy most of the data and a large number of tail classes only contain very limited samples. In practice, deep models often show poor generalization performance on tail classes due to the imbalanced distribution. To tackle this, data augmentation has become an effective way by synthesizing new samples for tail classes. Among them, one popular way is to use CutMix that explicitly mixups the images of tail classes and the others, while constructing the labels according to the ratio of areas cropped from two images. However, the area-based labels entirely ignore the inherent semantic information of the augmented samples, often leading to misleading training signals. To address this issue, we propose a Contrastive CutMix (ConCutMix) that constructs augmented samples with semantically consistent labels to boost the performance of long-tailed recognition. Specifically, we compute the similarities between samples in the semantic space learned by contrastive learning, and use them to rectify the area-based labels. Experiments show that our ConCutMix significantly improves the accuracy on tail classes as well as the overall performance. For example, based on ResNeXt-50, we improve the overall accuracy on ImageNet-LT by 3.0% thanks to the significant improvement of 3.3% on tail classes. We highlight that the improvement also generalizes well to other benchmarks and models. Our code and pretrained models are available at https://github.com/PanHaulin/ConCutMix.
Authors: Tianling Liu, Hongying Liu, Fanhua Shang, Lequan Yu, Tong Han, Liang Wan
Abstract: Multimodal MRIs play a crucial role in clinical diagnosis and treatment. Feature disentanglement (FD)-based methods, aiming at learning superior feature representations for multimodal data analysis, have achieved significant success in multimodal learning (MML). Typically, existing FD-based methods separate multimodal data into modality-shared and modality-specific features, and employ concatenation or attention mechanisms to integrate these features. However, our preliminary experiments indicate that these methods could lead to a loss of shared information among subsets of modalities when the inputs contain more than two modalities, and such information is critical for prediction accuracy. Furthermore, these methods do not adequately interpret the relationships between the decoupled features at the fusion stage. To address these limitations, we propose a novel Complete Feature Disentanglement (CFD) strategy that recovers the lost information during feature decoupling. Specifically, the CFD strategy not only identifies modality-shared and modality-specific features, but also decouples shared features among subsets of multimodal inputs, termed as modality-partial-shared features. We further introduce a new Dynamic Mixture-of-Experts Fusion (DMF) module that dynamically integrates these decoupled features, by explicitly learning the local-global relationships among the features. The effectiveness of our approach is validated through classification tasks on three multimodal MRI datasets. Extensive experimental results demonstrate that our approach outperforms other state-of-the-art MML methods with obvious margins, showcasing its superior performance.
Authors: Petrus Zwart
Abstract: In scientific imaging, deep learning has become a pivotal tool for image analytics. However, handling large volumetric datasets, which often exceed the memory capacity of standard GPUs, require special attention when subjected to deep learning efforts. This paper introduces qlty, a toolkit designed to address these challenges through tensor management techniques. qlty offers robust methods for subsampling, cleaning, and stitching of large-scale spatial data, enabling effective training and inference even in resource-limited environments.
Authors: Tsuyoshi Ishizone, Masaki Miyasaka, Sae Ochi, Norio Tada, Kazuyuki Nakamura
Abstract: Anatomical landmark localization is gaining attention to ease the burden on physicians. Focusing on aortic root landmark localization, the three hinge points of the aortic valve can reduce the burden by automatically determining the valve size required for transcatheter aortic valve implantation surgery. Existing methods for landmark prediction of the aortic root mainly use time-consuming two-step estimation methods. We propose a highly accurate one-step landmark localization method from even coarse images. The proposed method uses an optimal transport loss to break the trade-off between prediction precision and learning stability in conventional heatmap regression methods. We apply the proposed method to the 3D CT image dataset collected at Sendai Kousei Hospital and show that it significantly improves the estimation error over existing methods and other loss functions. Our code is available on GitHub.
Authors: Tiancheng Zhao, Qianqian Zhang, Kyusong Lee, Peng Liu, Lu Zhang, Chunxin Fang, Jiajia Liao, Kelei Jiang, Yibo Ma, Ruochen Xu
Abstract: We introduce OmChat, a model designed to excel in handling long contexts and video understanding tasks. OmChat's new architecture standardizes how different visual inputs are processed, making it more efficient and adaptable. It uses a dynamic vision encoding process to effectively handle images of various resolutions, capturing fine details across a range of image qualities. OmChat utilizes an active progressive multimodal pretraining strategy, which gradually increases the model's capacity for long contexts and enhances its overall abilities. By selecting high-quality data during training, OmChat learns from the most relevant and informative data points. With support for a context length of up to 512K, OmChat demonstrates promising performance in tasks involving multiple images and videos, outperforming most open-source models in these benchmarks. Additionally, OmChat proposes a prompting strategy for unifying complex multimodal inputs including single image text, multi-image text and videos, and achieving competitive performance on single-image benchmarks. To further evaluate the model's capabilities, we proposed a benchmark dataset named Temporal Visual Needle in a Haystack. This dataset assesses OmChat's ability to comprehend temporal visual details within long videos. Our analysis highlights several key factors contributing to OmChat's success: support for any-aspect high image resolution, the active progressive pretraining strategy, and high-quality supervised fine-tuning datasets. This report provides a detailed overview of OmChat's capabilities and the strategies that enhance its performance in visual understanding.
Authors: Brian Cheong, Jiachen Zhou, Steven Waslander
Abstract: Tracking-by-detection (TBD) methods achieve state-of-the-art performance on 3D tracking benchmarks for autonomous driving. On the other hand, tracking-by-attention (TBA) methods have the potential to outperform TBD methods, particularly for long occlusions and challenging detection settings. This work investigates why TBA methods continue to lag in performance behind TBD methods using a LiDAR-based joint detector and tracker called JDT3D. Based on this analysis, we propose two generalizable methods to bridge the gap between TBD and TBA methods: track sampling augmentation and confidence-based query propagation. JDT3D is trained and evaluated on the nuScenes dataset, achieving 0.574 on the AMOTA metric on the nuScenes test set, outperforming all existing LiDAR-based TBA approaches by over 6%. Based on our results, we further discuss some potential challenges with the existing TBA model formulation to explain the continued gap in performance with TBD methods. The implementation of JDT3D can be found at the following link: https://github.com/TRAILab/JDT3D.
Authors: Fengchuang Xing, Mingjie Li, Yuan-Gen Wang, Guopu Zhu, Xiaochun Cao
Abstract: In learning vision-language representations from web-scale data, the contrastive language-image pre-training (CLIP) mechanism has demonstrated a remarkable performance in many vision tasks. However, its application to the widely studied video quality assessment (VQA) task is still an open issue. In this paper, we propose an efficient and effective CLIP-based Transformer method for the VQA problem (CLIPVQA). Specifically, we first design an effective video frame perception paradigm with the goal of extracting the rich spatiotemporal quality and content information among video frames. Then, the spatiotemporal quality features are adequately integrated together using a self-attention mechanism to yield video-level quality representation. To utilize the quality language descriptions of videos for supervision, we develop a CLIP-based encoder for language embedding, which is then fully aggregated with the generated content information via a cross-attention module for producing video-language representation. Finally, the video-level quality and video-language representations are fused together for final video quality prediction, where a vectorized regression loss is employed for efficient end-to-end optimization. Comprehensive experiments are conducted on eight in-the-wild video datasets with diverse resolutions to evaluate the performance of CLIPVQA. The experimental results show that the proposed CLIPVQA achieves new state-of-the-art VQA performance and up to 37% better generalizability than existing benchmark VQA methods. A series of ablation studies are also performed to validate the effectiveness of each module in CLIPVQA.
Authors: Guoan Wang, Jin Ye, Junlong Cheng, Tianbin Li, Zhaolin Chen, Jianfei Cai, Junjun He, Bohan Zhuang
Abstract: Volumetric medical image segmentation is pivotal in enhancing disease diagnosis, treatment planning, and advancing medical research. While existing volumetric foundation models for medical image segmentation, such as SAM-Med3D and SegVol, have shown remarkable performance on general organs and tumors, their ability to segment certain categories in clinical downstream tasks remains limited. Supervised Finetuning (SFT) serves as an effective way to adapt such foundation models for task-specific downstream tasks but at the cost of degrading the general knowledge previously stored in the original foundation model.To address this, we propose SAM-Med3D-MoE, a novel framework that seamlessly integrates task-specific finetuned models with the foundational model, creating a unified model at minimal additional training expense for an extra gating network. This gating network, in conjunction with a selection strategy, allows the unified model to achieve comparable performance of the original models in their respective tasks both general and specialized without updating any parameters of them.Our comprehensive experiments demonstrate the efficacy of SAM-Med3D-MoE, with an average Dice performance increase from 53 to 56.4 on 15 specific classes. It especially gets remarkable gains of 29.6, 8.5, 11.2 on the spinal cord, esophagus, and right hip, respectively. Additionally, it achieves 48.9 Dice on the challenging SPPIN2023 Challenge, significantly surpassing the general expert's performance of 32.3. We anticipate that SAM-Med3D-MoE can serve as a new framework for adapting the foundation model to specific areas in medical image analysis. Codes and datasets will be publicly available.
Authors: Mohammadamin Baghbanbashi, Mohsen Raji, Behnam Ghavami
Abstract: YOLO is a deep neural network (DNN) model presented for robust real-time object detection following the one-stage inference approach. It outperforms other real-time object detectors in terms of speed and accuracy by a wide margin. Nevertheless, since YOLO is developed upon a DNN backbone with numerous parameters, it will cause excessive memory load, thereby deploying it on memory-constrained devices is a severe challenge in practice. To overcome this limitation, model compression techniques, such as quantizing parameters to lower-precision values, can be adopted. As the most recent version of YOLO, YOLOv7 achieves such state-of-the-art performance in speed and accuracy in the range of 5 FPS to 160 FPS that it surpasses all former versions of YOLO and other existing models in this regard. So far, the robustness of several quantization schemes has been evaluated on older versions of YOLO. These methods may not necessarily yield similar results for YOLOv7 as it utilizes a different architecture. In this paper, we conduct in-depth research on the effectiveness of a variety of quantization schemes on the pre-trained weights of the state-of-the-art YOLOv7 model. Experimental results demonstrate that using 4-bit quantization coupled with the combination of different granularities results in ~3.92x and ~3.86x memory-saving for uniform and non-uniform quantization, respectively, with only 2.5% and 1% accuracy loss compared to the full-precision baseline model.
Authors: Zhekai Chen, Wen Wang, Zhen Yang, Zeqing Yuan, Hao Chen, Chunhua Shen
Abstract: We offer a novel approach to image composition, which integrates multiple input images into a single, coherent image. Rather than concentrating on specific use cases such as appearance editing (image harmonization) or semantic editing (semantic image composition), we showcase the potential of utilizing the powerful generative prior inherent in large-scale pre-trained diffusion models to accomplish generic image composition applicable to both scenarios. We observe that the pre-trained diffusion models automatically identify simple copy-paste boundary areas as low-density regions during denoising. Building on this insight, we propose to optimize the composed image towards high-density regions guided by the diffusion prior. In addition, we introduce a novel maskguided loss to further enable flexible semantic image composition. Extensive experiments validate the superiority of our approach in achieving generic zero-shot image composition. Additionally, our approach shows promising potential in various tasks, such as object removal and multiconcept customization.
Authors: Huilin Zhu, Jingling Yuan, Zhengwei Yang, Yu Guo, Zheng Wang, Xian Zhong, Shengfeng He
Abstract: Zero-shot object counting (ZOC) aims to enumerate objects in images using only the names of object classes during testing, without the need for manual annotations. However, a critical challenge in current ZOC methods lies in their inability to identify high-quality exemplars effectively. This deficiency hampers scalability across diverse classes and undermines the development of strong visual associations between the identified classes and image content. To this end, we propose the Visual Association-based Zero-shot Object Counting (VA-Count) framework. VA-Count consists of an Exemplar Enhancement Module (EEM) and a Noise Suppression Module (NSM) that synergistically refine the process of class exemplar identification while minimizing the consequences of incorrect object identification. The EEM utilizes advanced vision-language pretaining models to discover potential exemplars, ensuring the framework's adaptability to various classes. Meanwhile, the NSM employs contrastive learning to differentiate between optimal and suboptimal exemplar pairs, reducing the negative effects of erroneous exemplars. VA-Count demonstrates its effectiveness and scalability in zero-shot contexts with superior performance on two object counting datasets.
Authors: Dingkang Yang, Mingcheng Li, Linhao Qu, Kun Yang, Peng Zhai, Song Wang, Lihua Zhang
Abstract: Understanding human intentions (e.g., emotions) from videos has received considerable attention recently. Video streams generally constitute a blend of temporal data stemming from distinct modalities, including natural language, facial expressions, and auditory clues. Despite the impressive advancements of previous works via attention-based paradigms, the inherent temporal asynchrony and modality heterogeneity challenges remain in multimodal sequence fusion, causing adverse performance bottlenecks. To tackle these issues, we propose a Multimodal fusion approach for learning modality-Exclusive and modality-Agnostic representations (MEA) to refine multimodal features and leverage the complementarity across distinct modalities. On the one hand, MEA introduces a predictive self-attention module to capture reliable context dynamics within modalities and reinforce unique features over the modality-exclusive spaces. On the other hand, a hierarchical cross-modal attention module is designed to explore valuable element correlations among modalities over the modality-agnostic space. Meanwhile, a double-discriminator strategy is presented to ensure the production of distinct representations in an adversarial manner. Eventually, we propose a decoupled graph fusion mechanism to enhance knowledge exchange across heterogeneous modalities and learn robust multimodal representations for downstream tasks. Numerous experiments are implemented on three multimodal datasets with asynchronous sequences. Systematic analyses show the necessity of our approach.
Authors: Dingkang Yang, Kun Yang, Haopeng Kuang, Zhaoyu Chen, Yuzheng Wang, Lihua Zhang
Abstract: Understanding emotions from diverse contexts has received widespread attention in computer vision communities. The core philosophy of Context-Aware Emotion Recognition (CAER) is to provide valuable semantic cues for recognizing the emotions of target persons by leveraging rich contextual information. Current approaches invariably focus on designing sophisticated structures to extract perceptually critical representations from contexts. Nevertheless, a long-neglected dilemma is that a severe context bias in existing datasets results in an unbalanced distribution of emotional states among different contexts, causing biased visual representation learning. From a causal demystification perspective, the harmful bias is identified as a confounder that misleads existing models to learn spurious correlations based on likelihood estimation, limiting the models' performance. To address the issue, we embrace causal inference to disentangle the models from the impact of such bias, and formulate the causalities among variables in the CAER task via a customized causal graph. Subsequently, we present a Contextual Causal Intervention Module (CCIM) to de-confound the confounder, which is built upon backdoor adjustment theory to facilitate seeking approximate causal effects during model training. As a plug-and-play component, CCIM can easily integrate with existing approaches and bring significant improvements. Systematic experiments on three datasets demonstrate the effectiveness of our CCIM.
Authors: Abhishek Sebastian, Annis Fathima A, Pragna R, Madhan Kumar S, Jesher Joshua M
Abstract: Community parks play a crucial role in promoting physical activity and overall well-being. This study introduces DLICP (Deep Learning Integrated Community Parks), an innovative approach that combines deep learning techniques specifically, face recognition technology with a novel walking activity measurement algorithm to enhance user experience in community parks. The DLICP utilizes a camera with face recognition software to accurately identify and track park users. Simultaneously, a walking activity measurement algorithm calculates parameters such as the average pace and calories burned, tailored to individual attributes. Extensive evaluations confirm the precision of DLICP, with a Mean Absolute Error (MAE) of 5.64 calories and a Mean Percentage Error (MPE) of 1.96%, benchmarked against widely available fitness measurement devices, such as the Apple Watch Series 6. This study contributes significantly to the development of intelligent smart park systems, enabling real-time updates on burned calories and personalized fitness tracking.
Authors: Haonan Xu, Dian Chao, Xiangyu Wu, Zhonghua Wan, Yang Yang
Abstract: Treating texts as images, combining prompts with textual labels for prompt tuning, and leveraging the alignment properties of CLIP have been successfully applied in zero-shot multi-label image recognition. Nonetheless, relying solely on textual labels to store visual information is insufficient for representing the diversity of visual objects. In this paper, we propose reversing the training process of CLIP and introducing the concept of Pseudo Visual Prompts. These prompts are initialized for each object category and pre-trained on large-scale, low-cost sentence data generated by large language models. This process mines the aligned visual information in CLIP and stores it in class-specific visual prompts. We then employ contrastive learning to transfer the stored visual information to the textual labels, enhancing their visual representation capacity. Additionally, we introduce a dual-adapter module that simultaneously leverages knowledge from the original CLIP and new learning knowledge derived from downstream datasets. Benefiting from the pseudo visual prompts, our method surpasses the state-of-the-art not only on clean annotated text data but also on pseudo text data generated by large language models.
Authors: Longfei Huang, Feng Yu, Zhihao Guan, Zhonghua Wan, Yang Yang
Abstract: This report presents a solution for the zero-shot referring expression comprehension task. Visual-language multimodal base models (such as CLIP, SAM) have gained significant attention in recent years as a cornerstone of mainstream research. One of the key applications of multimodal base models lies in their ability to generalize to zero-shot downstream tasks. Unlike traditional referring expression comprehension, zero-shot referring expression comprehension aims to apply pre-trained visual-language models directly to the task without specific training. Recent studies have enhanced the zero-shot performance of multimodal base models in referring expression comprehension tasks by introducing visual prompts. To address the zero-shot referring expression comprehension challenge, we introduced a combination of visual prompts and considered the influence of textual prompts, employing joint prediction tailored to the data characteristics. Ultimately, our approach achieved accuracy rates of 84.825 on the A leaderboard and 71.460 on the B leaderboard, securing the first position.
Authors: Vladyslav Polushko, Alexander Jenal, Jens Bongartz, Immanuel Weber, Damjan Hatic, Ronald R\"osch, Thomas M\"arz, Markus Rauhut, Andreas Weinmann
Abstract: Floods are an increasingly common global threat, causing emergencies and severe damage to infrastructure. During crises, organisations such as the World Food Programme use remotely sensed imagery, typically obtained through drones, for rapid situational analysis to plan life-saving actions. Computer Vision tools are needed to support task force experts on-site in the evaluation of the imagery to improve their efficiency and to allocate resources strategically. We introduce the BlessemFlood21 dataset to stimulate research on efficient flood detection tools. The imagery was acquired during the 2021 Erftstadt-Blessem flooding event and consists of high-resolution and georeferenced RGB-NIR images. In the resulting RGB dataset, the images are supplemented with detailed water masks, obtained via a semi-supervised human-in-the-loop technique, where in particular the NIR information is leveraged to classify pixels as either water or non-water. We evaluate our dataset by training and testing established Deep Learning models for semantic segmentation. With BlessemFlood21 we provide labeled high-resolution RGB data and a baseline for further development of algorithmic solutions tailored to flood detection in RGB imagery.
Authors: Fan Duan, Jiahao Yu, Li Chen
Abstract: Point clouds are commonly used in various practical applications such as autonomous driving and the manufacturing industry. However, these point clouds often suffer from incompleteness due to limited perspectives, scanner resolution and occlusion. Therefore the prediction of missing parts performs a crucial task. In this paper, we propose a novel method for point cloud completion. We utilize a spherical template to guide the generation of the coarse complete template and generate the dynamic query tokens through a correspondence pooling (Corres-Pooling) query generator. Specifically, we first generate the coarse complete template by embedding a Gaussian spherical template into the partial input and transforming the template to best match the input. Then we use the Corres-Pooling query generator to refine the coarse template and generate dynamic query tokens which could be used to predict the complete point proxies. Finally, we generate the complete point cloud with a FoldingNet following the coarse-to-fine paradigm, according to the fine template and the predicted point proxies. Experimental results demonstrate that our T-CorresNet outperforms the state-of-the-art methods on several benchmarks. Our Codes are available at https://github.com/df-boy/T-CorresNet.
Authors: Ye Li, Chen Tang, Yuan Meng, Jiajun Fan, Zenghao Chai, Xinzhu Ma, Zhi Wang, Wenwu Zhu
Abstract: We introduce PRANCE, a Vision Transformer compression framework that jointly optimizes the activated channels and reduces tokens, based on the characteristics of inputs. Specifically, PRANCE~ leverages adaptive token optimization strategies for a certain computational budget, aiming to accelerate ViTs' inference from a unified data and architectural perspective. However, the joint framework poses challenges to both architectural and decision-making aspects. Firstly, while ViTs inherently support variable-token inference, they do not facilitate dynamic computations for variable channels. To overcome this limitation, we propose a meta-network using weight-sharing techniques to support arbitrary channels of the Multi-head Self-Attention and Multi-layer Perceptron layers, serving as a foundational model for architectural decision-making. Second, simultaneously optimizing the structure of the meta-network and input data constitutes a combinatorial optimization problem with an extremely large decision space, reaching up to around $10^{14}$, making supervised learning infeasible. To this end, we design a lightweight selector employing Proximal Policy Optimization for efficient decision-making. Furthermore, we introduce a novel "Result-to-Go" training mechanism that models ViTs' inference process as a Markov decision process, significantly reducing action space and mitigating delayed-reward issues during training. Extensive experiments demonstrate the effectiveness of PRANCE~ in reducing FLOPs by approximately 50\%, retaining only about 10\% of tokens while achieving lossless Top-1 accuracy. Additionally, our framework is shown to be compatible with various token optimization techniques such as pruning, merging, and sequential pruning-merging strategies. The code is available at \href{https://github.com/ChildTang/PRANCE}{https://github.com/ChildTang/PRANCE}.
URLs: https://github.com/ChildTang/PRANCE, https://github.com/ChildTang/PRANCE
Authors: Xiaoya Cheng, Yu Liu, Maojun Zhang, Shen Yan
Abstract: In this paper, we present a novel approach for multiview point cloud registration. Different from previous researches that typically employ a global scheme for multiview registration, we propose to adopt an incremental pipeline to progressively align scans into a canonical coordinate system. Specifically, drawing inspiration from image-based 3D reconstruction, our approach first builds a sparse scan graph with scan retrieval and geometric verification. Then, we perform incremental registration via initialization, next scan selection and registration, Track create and continue, and Bundle Adjustment. Additionally, for detector-free matchers, we incorporate a Track refinement process. This process primarily constructs a coarse multiview registration and refines the model by adjusting the positions of the keypoints on the Track. Experiments demonstrate that the proposed framework outperforms existing multiview registration methods on three benchmark datasets. The code is available at https://github.com/Choyaa/IncreMVR.
Authors: Weixing Xie, Junfeng Yao, Xianpeng Cao, Qiqin Lin, Zerui Tang, Xiao Dong, Xiaohu Guo
Abstract: Dynamic reconstruction of deformable tissues in endoscopic video is a key technology for robot-assisted surgery. Recent reconstruction methods based on neural radiance fields (NeRFs) have achieved remarkable results in the reconstruction of surgical scenes. However, based on implicit representation, NeRFs struggle to capture the intricate details of objects in the scene and cannot achieve real-time rendering. In addition, restricted single view perception and occluded instruments also propose special challenges in surgical scene reconstruction. To address these issues, we develop SurgicalGaussian, a deformable 3D Gaussian Splatting method to model dynamic surgical scenes. Our approach models the spatio-temporal features of soft tissues at each time stamp via a forward-mapping deformation MLP and regularization to constrain local 3D Gaussians to comply with consistent movement. With the depth initialization strategy and tool mask-guided training, our method can remove surgical instruments and reconstruct high-fidelity surgical scenes. Through experiments on various surgical videos, our network outperforms existing method on many aspects, including rendering quality, rendering speed and GPU usage. The project page can be found at https://surgicalgaussian.github.io.
Authors: Dong Han, Yufan Jiang, Yong Li, Ricardo Mendes, Joachim Denzler
Abstract: In this work, we leverage the pure skin color patch from the face image as the additional information to train an auxiliary skin color feature extractor and face recognition model in parallel to improve performance of state-of-the-art (SOTA) privacy-preserving face recognition (PPFR) systems. Our solution is robust against black-box attacking and well-established generative adversarial network (GAN) based image restoration. We analyze the potential risk in previous work, where the proposed cosine similarity computation might directly leak the protected precomputed embedding stored on the server side. We propose a Function Secret Sharing (FSS) based face embedding comparison protocol without any intermediate result leakage. In addition, we show in experiments that the proposed protocol is more efficient compared to the Secret Sharing (SS) based protocol.
Authors: Monika Wysocza\'nska, Antonin Vobecky, Amaia Cardiel, Tomasz Trzci\'nski, Renaud Marlet, Andrei Bursuc, Oriane Sim\'eoni
Abstract: Recent VLMs, pre-trained on large amounts of image-text pairs to align both modalities, have opened the way to open-vocabulary semantic segmentation. Given an arbitrary set of textual queries, image regions are assigned the closest query in feature space. However, the usual setup expects the user to list all possible visual concepts that may occur in the image, typically all classes of benchmark datasets, that act as negatives to each other. We consider here the more challenging scenario of segmenting a single concept, given a textual prompt and nothing else. To achieve good results, besides contrasting with the generic 'background' text, we study different ways to generate query-specific test-time contrastive textual concepts, which leverage either the distribution of text in the VLM's training set or crafted LLM prompts. We show the relevance of our approach using a new, specific metric.
Authors: Kai Shen, Lingfei Wu, Siliang Tang, Fangli Xu, Bo Long, Yueting Zhuang, Jian Pei
Abstract: The visual question generation (VQG) task aims to generate human-like questions from an image and potentially other side information (e.g. answer type). Previous works on VQG fall in two aspects: i) They suffer from one image to many questions mapping problem, which leads to the failure of generating referential and meaningful questions from an image. ii) They fail to model complex implicit relations among the visual objects in an image and also overlook potential interactions between the side information and image. To address these limitations, we first propose a novel learning paradigm to generate visual questions with answer-awareness and region-reference. Concretely, we aim to ask the right visual questions with Double Hints - textual answers and visual regions of interests, which could effectively mitigate the existing one-to-many mapping issue. Particularly, we develop a simple methodology to self-learn the visual hints without introducing any additional human annotations. Furthermore, to capture these sophisticated relationships, we propose a new double-hints guided Graph-to-Sequence learning framework, which first models them as a dynamic graph and learns the implicit topology end-to-end, and then utilizes a graph-to-sequence model to generate the questions with double hints. Experimental results demonstrate the priority of our proposed method.
Authors: Qi Wang, Zhou Xu, Yuming Lin, Jingtao Ye, Hongsheng Li, Guangming Zhu, Syed Afaq Ali Shah, Mohammed Bennamoun, Liang Zhang
Abstract: Neuromorphic sensors, specifically event cameras, revolutionize visual data acquisition by capturing pixel intensity changes with exceptional dynamic range, minimal latency, and energy efficiency, setting them apart from conventional frame-based cameras. The distinctive capabilities of event cameras have ignited significant interest in the domain of event-based action recognition, recognizing their vast potential for advancement. However, the development in this field is currently slowed by the lack of comprehensive, large-scale datasets, which are critical for developing robust recognition frameworks. To bridge this gap, we introduces DailyDVS-200, a meticulously curated benchmark dataset tailored for the event-based action recognition community. DailyDVS-200 is extensive, covering 200 action categories across real-world scenarios, recorded by 47 participants, and comprises more than 22,000 event sequences. This dataset is designed to reflect a broad spectrum of action types, scene complexities, and data acquisition diversity. Each sequence in the dataset is annotated with 14 attributes, ensuring a detailed characterization of the recorded actions. Moreover, DailyDVS-200 is structured to facilitate a wide range of research paths, offering a solid foundation for both validating existing approaches and inspiring novel methodologies. By setting a new benchmark in the field, we challenge the current limitations of neuromorphic data processing and invite a surge of new approaches in event-based action recognition techniques, which paves the way for future explorations in neuromorphic computing and beyond. The dataset and source code are available at https://github.com/QiWang233/DailyDVS-200.
Authors: Zixu Cheng, Yujiang Pu, Shaogang Gong, Parisa Kordjamshidi, Yu Kong
Abstract: Temporal grounding, a.k.a video moment retrieval, aims at locating video segments corresponding to a given query sentence. The compositional nature of natural language enables the localization beyond predefined events, posing a certain challenge to the compositional generalizability of existing methods. Recent studies establish the correspondence between videos and queries through a decompose-reconstruct manner to achieve compositional generalization. However, they only consider dominant primitives and build negative queries through random sampling and recombination, resulting in semantically implausible negatives that hinder the models from learning rational compositions. In addition, recent DETR-based methods still underperform in compositional temporal grounding, showing irrational saliency responses when given negative queries that have subtle differences from positive queries. To address these limitations, we first propose a large language model-driven method for negative query construction, utilizing GPT-3.5-Turbo to generate semantically plausible hard negative queries. Subsequently, we introduce a coarse-to-fine saliency ranking strategy, which encourages the model to learn the multi-granularity semantic relationships between videos and hierarchical negative queries to boost compositional generalization. Extensive experiments on two challenging benchmarks validate the effectiveness and generalizability of our proposed method. Our code is available at https://github.com/zxccade/SHINE.
Authors: Yilu Wu, Hanlin Wang, Jing Wang, Limin Wang
Abstract: Given the current visual observations, the traditional procedure planning task in instructional videos requires a model to generate goal-directed plans within a given action space. All previous methods for this task conduct training and inference under the same action space, and they can only plan for pre-defined events in the training set. We argue this setting is not applicable for human assistance in real lives and aim to propose a more general and practical planning paradigm. Specifically, in this paper, we introduce a new task named Open-event Procedure Planning (OEPP), which extends the traditional procedure planning to the open-event setting. OEPP aims to verify whether a planner can transfer the learned knowledge to similar events that have not been seen during training. We rebuild a new benchmark of OpenEvent for this task based on existing datasets and divide the events involved into base and novel parts. During the data collection process, we carefully ensure the transfer ability of procedural knowledge for base and novel events by evaluating the similarity between the descriptions of different event steps with multiple stages. Based on the collected data, we further propose a simple and general framework specifically designed for OEPP, and conduct extensive study with various baseline methods, providing a detailed and insightful analysis on the results for this task.
Authors: Yunzhong Si, Huiying Xu, Xinzhong Zhu, Wenhao Zhang, Yao Dong, Yuxing Chen, Hongbo Li
Abstract: Channel and spatial attentions have respectively brought significant improvements in extracting feature dependencies and spatial structure relations for various downstream vision tasks. While their combination is more beneficial for leveraging their individual strengths, the synergy between channel and spatial attentions has not been fully explored, lacking in fully harness the synergistic potential of multi-semantic information for feature guidance and mitigation of semantic disparities. Our study attempts to reveal the synergistic relationship between spatial and channel attention at multiple semantic levels, proposing a novel Spatial and Channel Synergistic Attention module (SCSA). Our SCSA consists of two parts: the Shareable Multi-Semantic Spatial Attention (SMSA) and the Progressive Channel-wise Self-Attention (PCSA). SMSA integrates multi-semantic information and utilizes a progressive compression strategy to inject discriminative spatial priors into PCSA's channel self-attention, effectively guiding channel recalibration. Additionally, the robust feature interactions based on the self-attention mechanism in PCSA further mitigate the disparities in multi-semantic information among different sub-features within SMSA. We conduct extensive experiments on seven benchmark datasets, including classification on ImageNet-1K, object detection on MSCOCO 2017, segmentation on ADE20K, and four other complex scene detection datasets. Our results demonstrate that our proposed SCSA not only surpasses the current state-of-the-art attention but also exhibits enhanced generalization capabilities across various task scenarios. The code and models are available at: https://github.com/HZAI-ZJNU/SCSA.
Authors: Wei Dong, Han Zhou, Ruiyi Wang, Xiaohong Liu, Guangtao Zhai, Jun Chen
Abstract: Image dehazing, a pivotal task in low-level vision, aims to restore the visibility and detail from hazy images. Many deep learning methods with powerful representation learning capability demonstrate advanced performance on non-homogeneous dehazing, however, these methods usually struggle with processing high-resolution images (e.g., $4000 \times 6000$) due to their heavy computational demands. To address these challenges, we introduce an innovative non-homogeneous Dehazing method via Deformable Convolutional Transformer-like architecture (DehazeDCT). Specifically, we first design a transformer-like network based on deformable convolution v4, which offers long-range dependency and adaptive spatial aggregation capabilities and demonstrates faster convergence and forward speed. Furthermore, we leverage a lightweight Retinex-inspired transformer to achieve color correction and structure refinement. Extensive experiment results and highly competitive performance of our method in NTIRE 2024 Dense and Non-Homogeneous Dehazing Challenge, ranking second among all 16 submissions, demonstrate the superior capability of our proposed method. The code is available: https://github.com/movingforward100/Dehazing_R.
Authors: Julien Quarez, Matthew Elliot, Oscar Maccormac, Nawal Khan, Marc Modat, Sebastien Ourselin, Jonathan Shapey, Alejandro Granados
Abstract: In surgical skill assessment, Objective Structured Assessments of Technical Skills (OSATS scores) and the Global Rating Scale (GRS) are established tools for evaluating the performance of surgeons during training. These metrics, coupled with feedback on their performance, enable surgeons to improve and achieve standards of practice. Recent studies on the open-source dataset JIGSAW, which contains both GRS and OSATS labels, have focused on regressing GRS scores from kinematic signals, video data, or a combination of both. In this paper, we argue that regressing the GRS score, a unitless value, by itself is too restrictive, and variations throughout the surgical trial do not hold significant clinical meaning. To address this gap, we developed a recurrent transformer model that outputs the surgeon's performance throughout their training session by relating the model's hidden states to five OSATS scores derived from kinematic signals. These scores are averaged and aggregated to produce a GRS prediction, enabling assessment of the model's performance against the state-of-the-art (SOTA). We report Spearman's Correlation Coefficient (SCC), demonstrating that our model outperforms SOTA models for all tasks, except for Suturing under the leave-one-subject-out (LOSO) scheme (SCC 0.68-0.89), while achieving comparable performance for suturing and across tasks under the leave-one-user-out (LOUO) scheme (SCC 0.45-0.68) and beating SOTA for Needle Passing (0.69). We argue that relating final OSATS scores to short instances throughout a surgeon's procedure is more clinically meaningful than a single GRS score. This approach also allows us to translate quantitative predictions into qualitative feedback, which is crucial for any automated surgical skill assessment pipeline. A senior surgeon validated our model's behaviour and agreed with the semi-supervised predictions 77 \% (p = 0.006) of the time.
Authors: Huitong Pan, Qi Zhang, Cornelia Caragea, Eduard Dragut, Longin Jan Latecki
Abstract: Flowcharts are graphical tools for representing complex concepts in concise visual representations. This paper introduces the FlowLearn dataset, a resource tailored to enhance the understanding of flowcharts. FlowLearn contains complex scientific flowcharts and simulated flowcharts. The scientific subset contains 3,858 flowcharts sourced from scientific literature and the simulated subset contains 10,000 flowcharts created using a customizable script. The dataset is enriched with annotations for visual components, OCR, Mermaid code representation, and VQA question-answer pairs. Despite the proven capabilities of Large Vision-Language Models (LVLMs) in various visual understanding tasks, their effectiveness in decoding flowcharts - a crucial element of scientific communication - has yet to be thoroughly investigated. The FlowLearn test set is crafted to assess the performance of LVLMs in flowchart comprehension. Our study thoroughly evaluates state-of-the-art LVLMs, identifying existing limitations and establishing a foundation for future enhancements in this relatively underexplored domain. For instance, in tasks involving simulated flowcharts, GPT-4V achieved the highest accuracy (58%) in counting the number of nodes, while Claude recorded the highest accuracy (83%) in OCR tasks. Notably, no single model excels in all tasks within the FlowLearn framework, highlighting significant opportunities for further development.
Authors: Andrei Jarca, Florinel-Alin Croitoru, Radu Tudor Ionescu
Abstract: We propose Curriculum by Masking (CBM), a novel state-of-the-art curriculum learning strategy that effectively creates an easy-to-hard training schedule via patch (token) masking, offering significant accuracy improvements over the conventional training regime and previous curriculum learning (CL) methods. CBM leverages gradient magnitudes to prioritize the masking of salient image regions via a novel masking algorithm and a novel masking block. Our approach enables controlling sample difficulty via the patch masking ratio, generating an effective easy-to-hard curriculum by gradually introducing harder samples as training progresses. CBM operates with two easily configurable parameters, i.e. the number of patches and the curriculum schedule, making it a versatile curriculum learning approach for object recognition and detection. We conduct experiments with various neural architectures, ranging from convolutional networks to vision transformers, on five benchmark data sets (CIFAR-10, CIFAR-100, ImageNet, Food-101 and PASCAL VOC), to compare CBM with conventional as well as curriculum-based training regimes. Our results reveal the superiority of our strategy compared with the state-of-the-art curriculum learning regimes. We also observe improvements in transfer learning contexts, where CBM surpasses previous work by considerable margins in terms of accuracy. We release our code for free non-commercial use at https://github.com/CroitoruAlin/CBM.
Authors: Prarthana Bhattacharyya, Joshua Mitton, Ryan Page, Owen Morgan, Ben Menzies, Gabriel Homewood, Kemi Jacobs, Paolo Baesso, Dave Trickett, Chris Mair, Taru Muhonen, Rory Clark, Louis Berridge, Richard Vigars, Iain Wallace
Abstract: This paper introduces Helios, the first extremely low-power, real-time, event-based hand gesture recognition system designed for all-day on smart eyewear. As augmented reality (AR) evolves, current smart glasses like the Meta Ray-Bans prioritize visual and wearable comfort at the expense of functionality. Existing human-machine interfaces (HMIs) in these devices, such as capacitive touch and voice controls, present limitations in ergonomics, privacy and power consumption. Helios addresses these challenges by leveraging natural hand interactions for a more intuitive and comfortable user experience. Our system utilizes a extremely low-power and compact 3mmx4mm/20mW event camera to perform natural hand-based gesture recognition for always-on smart eyewear. The camera's output is processed by a convolutional neural network (CNN) running on a NXP Nano UltraLite compute platform, consuming less than 350mW. Helios can recognize seven classes of gestures, including subtle microgestures like swipes and pinches, with 91% accuracy. We also demonstrate real-time performance across 20 users at a remarkably low latency of 60ms. Our user testing results align with the positive feedback we received during our recent successful demo at AWE-USA-2024.
Authors: Harshkumar Devmurari, Gautham Kuckian, Prajjwal Vishwakarma, Krunali Vartak
Abstract: Generating images from hand-drawings is a crucial and fundamental task in content creation. The translation is challenging due to the infinite possibilities and the diverse expectations of users. However, traditional methods are often limited by the availability of training data. Therefore, VisioBlend, a unified framework supporting three-dimensional control over image synthesis from sketches and strokes based on diffusion models, is proposed. It enables users to decide the level of faithfulness to the input strokes and sketches. VisioBlend achieves state-of-the-art performance in terms of realism and flexibility, enabling various applications in image synthesis from sketches and strokes. It solves the problem of data availability by synthesizing new data points from hand-drawn sketches and strokes, enriching the dataset and enabling more robust and diverse image synthesis. This work showcases the power of diffusion models in image creation, offering a user-friendly and versatile approach for turning artistic visions into reality.
Authors: Amy Saranchuk, Michael Guerzhoy
Abstract: Self-supervised learning for pre-training (SSP) can help the network learn better low-level features, especially when the size of the training set is small. In contrastive pre-training, the network is pre-trained to distinguish between different versions of the input. For example, the network learns to distinguish pairs (original, rotated) of images where the rotated image was rotated by angle $\theta$ vs. other pairs of images. In this work, we show that, when training using contrastive pre-training in this way, the angle $\theta$ and the dataset interact in interesting ways. We hypothesize, and give some evidence, that, for some datasets, the network can take "shortcuts" for particular rotation angles $\theta$ based on the distribution of the gradient directions in the input, possibly avoiding learning features other than edges, but our experiments do not seem to support that hypothesis. We demonstrate experiments on three radiology datasets. We compute the saliency map indicating which pixels were important in the SSP process, and compare the saliency map to the ground truth foreground/background segmentation. Our visualizations indicate that the effects of rotation angles in SSP are dataset-dependent. We believe the distribution of gradient orientations may play a role in this, but our experiments so far are inconclusive.
Authors: Xiaoyu Guo, Pengzhi Zhong, Lizhi Lin, Hao Zhang, Ling Huang, Shuiwang Li
Abstract: Visual tracking has advanced significantly in recent years, mainly due to the availability of large-scale training datasets. These datasets have enabled the development of numerous algorithms that can track objects with high accuracy and robustness.However, the majority of current research has been directed towards tracking generic objects, with less emphasis on more specialized and challenging scenarios. One such challenging scenario involves tracking reflected objects. Reflections can significantly distort the appearance of objects, creating ambiguous visual cues that complicate the tracking process. This issue is particularly pertinent in applications such as autonomous driving, security, smart homes, and industrial production, where accurately tracking objects reflected in surfaces like mirrors or glass is crucial. To address this gap, we introduce TRO, a benchmark specifically for Tracking Reflected Objects. TRO includes 200 sequences with around 70,000 frames, each carefully annotated with bounding boxes. This dataset aims to encourage the development of new, accurate methods for tracking reflected objects, which present unique challenges not sufficiently covered by existing benchmarks. We evaluated 20 state-of-the-art trackers and found that they struggle with the complexities of reflections. To provide a stronger baseline, we propose a new tracker, HiP-HaTrack, which uses hierarchical features to improve performance, significantly outperforming existing algorithms. We believe our benchmark, evaluation, and HiP-HaTrack will inspire further research and applications in tracking reflected objects. The TRO and code are available at https://github.com/OpenCodeGithub/HIP-HaTrack.
Authors: Jiahao Nie, Fei Xie, Xueyi Zhou, Sifan Zhou, Zhiwei He, Dong-Kyu Chae
Abstract: 3D single object tracking (SOT) methods based on appearance matching has long suffered from insufficient appearance information incurred by incomplete, textureless and semantically deficient LiDAR point clouds. While motion paradigm exploits motion cues instead of appearance matching for tracking, it incurs complex multi-stage processing and segmentation module. In this paper, we first provide in-depth explorations on motion paradigm, which proves that (\textbf{i}) it is feasible to directly infer target relative motion from point clouds across consecutive frames; (\textbf{ii}) fine-grained information comparison between consecutive point clouds facilitates target motion modeling. We thereby propose to perform part-to-part motion modeling for consecutive point clouds and introduce a novel tracking framework, termed \textbf{P2P}. The novel framework fuses each corresponding part information between consecutive point clouds, effectively exploring detailed information changes and thus modeling accurate target-related motion cues. Following this framework, we present P2P-point and P2P-voxel models, incorporating implicit and explicit part-to-part motion modeling by point- and voxel-based representation, respectively. Without bells and whistles, P2P-voxel sets a new state-of-the-art performance ($\sim$\textbf{89\%}, \textbf{72\%} and \textbf{63\%} precision on KITTI, NuScenes and Waymo Open Dataset, respectively). Moreover, under the same point-based representation, P2P-point outperforms the previous motion tracker M$^2$Track by \textbf{3.3\%} and \textbf{6.7\%} on the KITTI and NuScenes, while running at a considerably high speed of \textbf{107 Fps} on a single RTX3090 GPU. The source code and pre-trained models are available at \url{https://github.com/haooozi/P2P}.
Authors: Junming Su, Zhiqiang Shen, Peng Cao, Jinzhu Yang, Osmar R. Zaiane
Abstract: The existing barely-supervised medical image segmentation (BSS) methods, adopting a registration-segmentation paradigm, aim to learn from data with very few annotations to mitigate the extreme label scarcity problem. However, this paradigm poses a challenge: pseudo-labels generated by image registration come with significant noise. To address this issue, we propose a self-paced sample selection framework (SPSS) for BSS. Specifically, SPSS comprises two main components: 1) self-paced uncertainty sample selection (SU) for explicitly improving the quality of pseudo labels in the image space, and 2) self-paced bidirectional feature contrastive learning (SC) for implicitly improving the quality of pseudo labels through enhancing the separability between class semantics in the feature space. Both SU and SC are trained collaboratively in a self-paced learning manner, ensuring that SPSS can learn from high-quality pseudo labels for BSS. Extensive experiments on two public medical image segmentation datasets demonstrate the effectiveness and superiority of SPSS over the state-of-the-art. Our code is release at https://github.com/SuuuJM/SPSS.
Authors: Jiahao Chang, Yinglin Xu, Yihao Li, Yuantao Chen, Xiaoguang Han
Abstract: Point cloud registration is a fundamental problem for large-scale 3D scene scanning and reconstruction. With the help of deep learning, registration methods have evolved significantly, reaching a nearly-mature stage. As the introduction of Neural Radiance Fields (NeRF), it has become the most popular 3D scene representation as its powerful view synthesis capabilities. Regarding NeRF representation, its registration is also required for large-scale scene reconstruction. However, this topic extremly lacks exploration. This is due to the inherent challenge to model the geometric relationship among two scenes with implicit representations. The existing methods usually convert the implicit representation to explicit representation for further registration. Most recently, Gaussian Splatting (GS) is introduced, employing explicit 3D Gaussian. This method significantly enhances rendering speed while maintaining high rendering quality. Given two scenes with explicit GS representations, in this work, we explore the 3D registration task between them. To this end, we propose GaussReg, a novel coarse-to-fine framework, both fast and accurate. The coarse stage follows existing point cloud registration methods and estimates a rough alignment for point clouds from GS. We further newly present an image-guided fine registration approach, which renders images from GS to provide more detailed geometric information for precise alignment. To support comprehensive evaluation, we carefully build a scene-level dataset called ScanNet-GSReg with 1379 scenes obtained from the ScanNet dataset and collect an in-the-wild dataset called GSReg. Experimental results demonstrate our method achieves state-of-the-art performance on multiple datasets. Our GaussReg is 44 times faster than HLoc (SuperPoint as the feature extractor and SuperGlue as the matcher) with comparable accuracy.
Authors: Shikha Verma, Kuldeep Srivastava, Akhilesh Tiwari, Shekhar Verma
Abstract: The rainfall associated with Topical Cyclone(TC) contributes a major amount to the annual rainfall in India. Due to the limited research on the quantitative precipitation associated with Tropical Cyclones (TC), the prediction of the amount of precipitation and area that it may cover remains a challenge. This paper proposes an approach to estimate the accumulated precipitation and impact on affected area using Remote Sensing data. For this study, an instance of Extremely Severe Cyclonic Storm, Biparjoy that formed over the Arabian Sea and hit India in 2023 is considered in which we have used the satellite images of IMERG-Late Run of Global Precipitation Measurement (GPM). Image processing techniques were employed to identify and extract precipitation clusters linked to the cyclone. The results indicate that Biparjoy contributed a daily average rainfall of 53.14 mm/day across India and the Arabian Sea, with the Indian boundary receiving 11.59 mm/day, covering an extensive 411.76 thousand square kilometers. The localized intensity and variability observed in states like Gujarat, Rajasthan, Madhya Pradesh, and Uttar Pradesh highlight the need for tailored response measures, emphasizing the importance of further research to enhance predictive models and disaster readiness, crucial for building resilience against the diverse impacts of tropical cyclones.
Authors: Pengkun Jiao, Na Zhao, Jingjing Chen, Yu-Gang Jiang
Abstract: Open-vocabulary 3D object detection (OV-3DDet) aims to localize and recognize both seen and previously unseen object categories within any new 3D scene. While language and vision foundation models have achieved success in handling various open-vocabulary tasks with abundant training data, OV-3DDet faces a significant challenge due to the limited availability of training data. Although some pioneering efforts have integrated vision-language models (VLM) knowledge into OV-3DDet learning, the full potential of these foundational models has yet to be fully exploited. In this paper, we unlock the textual and visual wisdom to tackle the open-vocabulary 3D detection task by leveraging the language and vision foundation models. We leverage a vision foundation model to provide image-wise guidance for discovering novel classes in 3D scenes. Specifically, we utilize a object detection vision foundation model to enable the zero-shot discovery of objects in images, which serves as the initial seeds and filtering guidance to identify novel 3D objects. Additionally, to align the 3D space with the powerful vision-language space, we introduce a hierarchical alignment approach, where the 3D feature space is aligned with the vision-language feature space using a pre-trained VLM at the instance, category, and scene levels. Through extensive experimentation, we demonstrate significant improvements in accuracy and generalization, highlighting the potential of foundation models in advancing open-vocabulary 3D object detection in real-world scenarios.
Authors: Jingyang Xiang, Zuohui Chen, Siqi Li, Qing Wu, Yong Liu
Abstract: Binary Neural Networks~(BNNs) have been proven to be highly effective for deploying deep neural networks on mobile and embedded platforms. Most existing works focus on minimizing quantization errors, improving representation ability, or designing gradient approximations to alleviate gradient mismatch in BNNs, while leaving the weight sign flipping, a critical factor for achieving powerful BNNs, untouched. In this paper, we investigate the efficiency of weight sign updates in BNNs. We observe that, for vanilla BNNs, over 50\% of the weights remain their signs unchanged during training, and these weights are not only distributed at the tails of the weight distribution but also universally present in the vicinity of zero. We refer to these weights as ``silent weights'', which slow down convergence and lead to a significant accuracy degradation. Theoretically, we reveal this is due to the independence of the BNNs gradient from the latent weight distribution. To address the issue, we propose Overcome Silent Weights~(OvSW). OvSW first employs Adaptive Gradient Scaling~(AGS) to establish a relationship between the gradient and the latent weight distribution, thereby improving the overall efficiency of weight sign updates. Additionally, we design Silence Awareness Decaying~(SAD) to automatically identify ``silent weights'' by tracking weight flipping state, and apply an additional penalty to ``silent weights'' to facilitate their flipping. By efficiently updating weight signs, our method achieves faster convergence and state-of-the-art performance on CIFAR10 and ImageNet1K dataset with various architectures. For example, OvSW obtains 61.6\% and 65.5\% top-1 accuracy on the ImageNet1K using binarized ResNet18 and ResNet34 architecture respectively. Codes are available at \url{https://github.com/JingyangXiang/OvSW}.
Authors: Akshat Ramachandran, Souvik Kundu, Tushar Krishna
Abstract: We present CLAMP-ViT, a data-free post-training quantization method for vision transformers (ViTs). We identify the limitations of recent techniques, notably their inability to leverage meaningful inter-patch relationships, leading to the generation of simplistic and semantically vague data, impacting quantization accuracy. CLAMP-ViT employs a two-stage approach, cyclically adapting between data generation and model quantization. Specifically, we incorporate a patch-level contrastive learning scheme to generate richer, semantically meaningful data. Furthermore, we leverage contrastive learning in layer-wise evolutionary search for fixed- and mixed-precision quantization to identify optimal quantization parameters while mitigating the effects of a non-smooth loss landscape. Extensive evaluations across various vision tasks demonstrate the superiority of CLAMP-ViT, with performance improvements of up to 3% in top-1 accuracy for classification, 0.6 mAP for object detection, and 1.5 mIoU for segmentation at similar or better compression ratio over existing alternatives. Code is available at https://github.com/georgia-tech-synergy-lab/CLAMP-ViT.git
URLs: https://github.com/georgia-tech-synergy-lab/CLAMP-ViT.git
Authors: Ting-Wei Zhou, Xi-Le Zhao, Jian-Li Wang, Yi-Si Luo, Min Wang, Xiao-Xuan Bai, Hong Yan
Abstract: Recently, the transform-based tensor representation has attracted increasing attention in multimedia data (e.g., images and videos) recovery problems, which consists of two indispensable components, i.e., transform and characterization. Previously, the development of transform-based tensor representation mainly focuses on the transform aspect. Although several attempts consider using shallow matrix factorization (e.g., singular value decomposition and negative matrix factorization) to characterize the frontal slices of transformed tensor (termed as latent tensor), the faithful characterization aspect is underexplored. To address this issue, we propose a unified Deep Tensor Representation (termed as DTR) framework by synergistically combining the deep latent generative module and the deep transform module. Especially, the deep latent generative module can faithfully generate the latent tensor as compared with shallow matrix factorization. The new DTR framework not only allows us to better understand the classic shallow representations, but also leads us to explore new representation. To examine the representation ability of the proposed DTR, we consider the representative multi-dimensional data recovery task and suggest an unsupervised DTR-based multi-dimensional data recovery model. Extensive experiments demonstrate that DTR achieves superior performance compared to state-of-the-art methods in both quantitative and qualitative aspects, especially for fine details recovery.
Authors: Valeriy Lobanov, Nikita Firsov, Evgeny Myasnikov, Roman Khabibullin, Artem Nikonorov
Abstract: In traditional neural network architectures, a multilayer perceptron (MLP) is typically employed as a classification block following the feature extraction stage. However, the Kolmogorov-Arnold Network (KAN) presents a promising alternative to MLP, offering the potential to enhance prediction accuracy. In this paper, we propose the replacement of linear and convolutional layers of traditional networks with KAN-based counterparts. These modifications allowed us to significantly increase the per-pixel classification accuracy for hyperspectral remote-sensing images. We modified seven different neural network architectures for hyperspectral image classification and observed a substantial improvement in the classification accuracy across all the networks. The architectures considered in the paper include baseline MLP, state-of-the-art 1D (1DCNN) and 3D convolutional (two different 3DCNN, NM3DCNN), and transformer (SSFTT) architectures, as well as newly proposed M1DCNN. The greatest effect was achieved for convolutional networks working exclusively on spectral data, and the best classification quality was achieved using a KAN-based transformer architecture. All the experiments were conducted using seven openly available hyperspectral datasets. Our code is available at https://github.com/f-neumann77/HyperKAN.
Authors: Kunjing Yang, Minru Bai, TingLu
Abstract: Fusing hyperspectral images (HSIs) with multispectral images (MSIs) has become a mainstream approach to enhance the spatial resolution of HSIs. Many HSI-MSI fusion methods have achieved impressive results. Nevertheless, certain challenges persist, including: (a) A majority of current methods rely on accurate registration of HSI and MSI, which can be challenging in real-world applications.(b) The obtained HSI-MSI pairs may not be fully utilized. In this paper, we propose a hybrid registration and fusion constrained optimization model named RAF-NLRGS. With respect to challenge (a), the RAF model integrates batch image alignment within the fusion process, facilitating simultaneous execution of image registration and fusion. To address issue (b), the NLRGS model incorporates a nonconvex low-rank and group-sparse structure, leveraging group sparsity to effectively harness valuable information embedded in the residual data. Moreover, the NLRGS model can further enhance fusion performance based on the RAF model. Subsequently, the RAF-NLRGS model is solved within the framework of Generalized Gauss-Newton (GGN) algorithm and Proximal Alternating Optimization (PAO) algorithm. Theoretically, we establish the error bounds for the NLRGS model and the convergence analysis of corresponding algorithms is also presented. Finally, extensive numerical experiments on HSI datasets are conducted to verify the effectiveness of our method.
Authors: Haozhe Zhao, Xiaojian Ma, Liang Chen, Shuzheng Si, Rujie Wu, Kaikai An, Peiyu Yu, Minjia Zhang, Qing Li, Baobao Chang
Abstract: This paper presents UltraEdit, a large-scale (approximately 4 million editing samples), automatically generated dataset for instruction-based image editing. Our key idea is to address the drawbacks in existing image editing datasets like InstructPix2Pix and MagicBrush, and provide a systematic approach to producing massive and high-quality image editing samples. UltraEdit offers several distinct advantages: 1) It features a broader range of editing instructions by leveraging the creativity of large language models (LLMs) alongside in-context editing examples from human raters; 2) Its data sources are based on real images, including photographs and artworks, which provide greater diversity and reduced bias compared to datasets solely generated by text-to-image models; 3) It also supports region-based editing, enhanced by high-quality, automatically produced region annotations. Our experiments show that canonical diffusion-based editing baselines trained on UltraEdit set new records on MagicBrush and Emu-Edit benchmarks. Our analysis further confirms the crucial role of real image anchors and region-based editing data. The dataset, code, and models can be found in https://ultra-editing.github.io.
Authors: Yi Feng, Zizhan Guo, Qijun Chen, Rui Fan
Abstract: Unsupervised monocular depth estimation frameworks have shown promising performance in autonomous driving. However, existing solutions primarily rely on a simple convolutional neural network for ego-motion recovery, which struggles to estimate precise camera poses in dynamic, complicated real-world scenarios. These inaccurately estimated camera poses can inevitably deteriorate the photometric reconstruction and mislead the depth estimation networks with wrong supervisory signals. In this article, we introduce SCIPaD, a novel approach that incorporates spatial clues for unsupervised depth-pose joint learning. Specifically, a confidence-aware feature flow estimator is proposed to acquire 2D feature positional translations and their associated confidence levels. Meanwhile, we introduce a positional clue aggregator, which integrates pseudo 3D point clouds from DepthNet and 2D feature flows into homogeneous positional representations. Finally, a hierarchical positional embedding injector is proposed to selectively inject spatial clues into semantic features for robust camera pose decoding. Extensive experiments and analyses demonstrate the superior performance of our model compared to other state-of-the-art methods. Remarkably, SCIPaD achieves a reduction of 22.2\% in average translation error and 34.8\% in average angular error for camera pose estimation task on the KITTI Odometry dataset. Our source code is available at \url{https://mias.group/SCIPaD}.
Authors: Kun Li, Dan Guo, Pengyu Liu, Guoliang Chen, Meng Wang
Abstract: Human body actions are an important form of non-verbal communication in social interactions. This paper focuses on a specific subset of body actions known as micro-actions, which are subtle, low-intensity body movements that provide a deeper understanding of inner human feelings. In real-world scenarios, human micro-actions often co-occur, with multiple micro-actions overlapping in time, such as simultaneous head and hand movements. However, current research primarily focuses on recognizing individual micro-actions while overlooking their co-occurring nature. To narrow this gap, we propose a new task named Multi-label Micro-Action Detection (MMAD), which involves identifying all micro-actions in a given short video, determining their start and end times, and categorizing them. Achieving this requires a model capable of accurately capturing both long-term and short-term action relationships to locate and classify multiple micro-actions. To support the MMAD task, we introduce a new dataset named Multi-label Micro-Action-52 (MMA-52), specifically designed to facilitate the detailed analysis and exploration of complex human micro-actions. The proposed MMA-52 dataset is available at: https://github.com/VUT-HFUT/Micro-Action.
Authors: Yan Zeng, Masanori Suganuma, Takayuki Okatani
Abstract: Diffusion models have demonstrated impressive image generation capabilities. Personalized approaches, such as textual inversion and Dreambooth, enhance model individualization using specific images. These methods enable generating images of specific objects based on diverse textual contexts. Our proposed approach aims to retain the model's original knowledge during new information integration, resulting in superior outcomes while necessitating less training time compared to Dreambooth and textual inversion.
Authors: Eun Som Jeon, Rahul Khurana, Aishani Pathak, Pavan Turaga
Abstract: Deep learning has shown its efficacy in extracting useful features to solve various computer vision tasks. However, when the structure of the data is complex and noisy, capturing effective information to improve performance is very difficult. To this end, topological data analysis (TDA) has been utilized to derive useful representations that can contribute to improving performance and robustness against perturbations. Despite its effectiveness, the requirements for large computational resources and significant time consumption in extracting topological features through TDA are critical problems when implementing it on small devices. To address this issue, we propose a framework called Topological Guidance-based Knowledge Distillation (TGD), which uses topological features in knowledge distillation (KD) for image classification tasks. We utilize KD to train a superior lightweight model and provide topological features with multiple teachers simultaneously. We introduce a mechanism for integrating features from different teachers and reducing the knowledge gap between teachers and the student, which aids in improving performance. We demonstrate the effectiveness of our approach through diverse empirical evaluations.
Authors: Bo Peng, Yunfan Tao, Haoyu Zhan, Yudong Guo, Juyong Zhang
Abstract: We introduce PICA, a novel representation for high-fidelity animatable clothed human avatars with physics-accurate dynamics, even for loose clothing. Previous neural rendering-based representations of animatable clothed humans typically employ a single model to represent both the clothing and the underlying body. While efficient, these approaches often fail to accurately represent complex garment dynamics, leading to incorrect deformations and noticeable rendering artifacts, especially for sliding or loose garments. Furthermore, previous works represent garment dynamics as pose-dependent deformations and facilitate novel pose animations in a data-driven manner. This often results in outcomes that do not faithfully represent the mechanics of motion and are prone to generating artifacts in out-of-distribution poses. To address these issues, we adopt two individual 3D Gaussian Splatting (3DGS) models with different deformation characteristics, modeling the human body and clothing separately. This distinction allows for better handling of their respective motion characteristics. With this representation, we integrate a graph neural network (GNN)-based clothed body physics simulation module to ensure an accurate representation of clothing dynamics. Our method, through its carefully designed features, achieves high-fidelity rendering of clothed human bodies in complex and novel driving poses, significantly outperforming previous methods under the same settings.
Authors: Andr\'e Longon
Abstract: A mechanistic understanding of the computations learned by deep neural networks (DNNs) is far from complete. In the domain of visual object recognition, prior research has illuminated inner workings of InceptionV1, but DNNs with different architectures have remained largely unexplored. This work investigates ResNet18 with a particular focus on its residual stream, an architectural mechanism which InceptionV1 lacks. We observe that for a given block, channel features of the stream are updated along a spectrum: either the input feature skips to the output, the block feature overwrites the output, or the output is some mixture between the input and block features. Furthermore, we show that many residual stream channels compute scale invariant representations through a mixture of the input's smaller-scale feature with the block's larger-scale feature. This not only mounts evidence for the universality of scale equivariance, but also presents how the residual stream further implements scale invariance. Collectively, our results begin an interpretation of the residual stream in visual object recognition, finding it to be a flexible feature manager and a medium to build scale invariant representations.
Authors: Longxiang Tang, Zhuotao Tian, Kai Li, Chunming He, Hantao Zhou, Hengshuang Zhao, Xiu Li, Jiaya Jia
Abstract: This study addresses the Domain-Class Incremental Learning problem, a realistic but challenging continual learning scenario where both the domain distribution and target classes vary across tasks. To handle these diverse tasks, pre-trained Vision-Language Models (VLMs) are introduced for their strong generalizability. However, this incurs a new problem: the knowledge encoded in the pre-trained VLMs may be disturbed when adapting to new tasks, compromising their inherent zero-shot ability. Existing methods tackle it by tuning VLMs with knowledge distillation on extra datasets, which demands heavy computation overhead. To address this problem efficiently, we propose the Distribution-aware Interference-free Knowledge Integration (DIKI) framework, retaining pre-trained knowledge of VLMs from a perspective of avoiding information interference. Specifically, we design a fully residual mechanism to infuse newly learned knowledge into a frozen backbone, while introducing minimal adverse impacts on pre-trained knowledge. Besides, this residual property enables our distribution-aware integration calibration scheme, explicitly controlling the information implantation process for test data from unseen distributions. Experiments demonstrate that our DIKI surpasses the current state-of-the-art approach using only 0.86% of the trained parameters and requiring substantially less training time. Code is available at: https://github.com/lloongx/DIKI .
Authors: Danni Yang, Ruohan Dong, Jiayi Ji, Yiwei Ma, Haowei Wang, Xiaoshuai Sun, Rongrong Ji
Abstract: Recently, diffusion models have increasingly demonstrated their capabilities in vision understanding. By leveraging prompt-based learning to construct sentences, these models have shown proficiency in classification and visual grounding tasks. However, existing approaches primarily showcase their ability to perform sentence-level localization, leaving the potential for leveraging contextual information for phrase-level understanding largely unexplored. In this paper, we utilize Panoptic Narrative Grounding (PNG) as a proxy task to investigate this capability further. PNG aims to segment object instances mentioned by multiple noun phrases within a given narrative text. Specifically, we introduce the DiffPNG framework, a straightforward yet effective approach that fully capitalizes on the diffusion's architecture for segmentation by decomposing the process into a sequence of localization, segmentation, and refinement steps. The framework initially identifies anchor points using cross-attention mechanisms and subsequently performs segmentation with self-attention to achieve zero-shot PNG. Moreover, we introduce a refinement module based on SAM to enhance the quality of the segmentation masks. Our extensive experiments on the PNG dataset demonstrate that DiffPNG achieves strong performance in the zero-shot PNG task setting, conclusively proving the diffusion model's capability for context-aware, phrase-level understanding. Source code is available at \url{https://github.com/nini0919/DiffPNG}.
Authors: Yan Wang, Yawen Zeng, Jingsheng Zheng, Xiaofen Xing, Jin Xu, Xiangmin Xu
Abstract: Multimodal large language models (MLLMs) are flourishing, but mainly focus on images with less attention than videos, especially in sub-fields such as prompt engineering, video chain-of-thought (CoT), and instruction tuning on videos. Therefore, we try to explore the collection of CoT datasets in videos to lead to video OpenQA and improve the reasoning ability of MLLMs. Unfortunately, making such video CoT datasets is not an easy task. Given that human annotation is too cumbersome and expensive, while machine-generated is not reliable due to the hallucination issue, we develop an automatic annotation tool that combines machine and human experts, under the active learning paradigm. Active learning is an interactive strategy between the model and human experts, in this way, the workload of human labeling can be reduced and the quality of the dataset can be guaranteed. With the help of the automatic annotation tool, we strive to contribute three datasets, namely VideoCoT, TopicQA, TopicCoT. Furthermore, we propose a simple but effective benchmark based on the collected datasets, which exploits CoT to maximize the complex reasoning capabilities of MLLMs. Extensive experiments demonstrate the effectiveness our solution.
Authors: Michael Welter
Abstract: Deep learning has been impressively successful in the last decade in predicting human head poses from monocular images. For in-the-wild inputs, the research community has predominantly relied on a single training set of semi-synthetic nature. This paper suggest the combination of different flavors of synthetic data in order to achieve better generalization to natural images. Moreover, additional expansion of the data volume using traditional out-of-plane rotation synthesis is considered. Together with a novel combination of losses and a network architecture with a standard feature-extractor, a competitive model is obtained, both in accuracy and efficiency, which allows full 6 DoF pose estimation in practical real-time applications.
Authors: Yuanhong Chen, Chong Wang, Yuyuan Liu, Hu Wang, Gustavo Carneiro
Abstract: Audio-visual segmentation (AVS) is an emerging task that aims to accurately segment sounding objects based on audio-visual cues. The success of AVS learning systems depends on the effectiveness of cross-modal interaction. Such a requirement can be naturally fulfilled by leveraging transformer-based segmentation architecture due to its inherent ability to capture long-range dependencies and flexibility in handling different modalities. However, the inherent training issues of transformer-based methods, such as the low efficacy of cross-attention and unstable bipartite matching, can be amplified in AVS, particularly when the learned audio query does not provide a clear semantic clue. In this paper, we address these two issues with the new Class-conditional Prompting Machine (CPM). CPM improves the bipartite matching with a learning strategy combining class-agnostic queries with class-conditional queries. The efficacy of cross-modal attention is upgraded with new learning objectives for the audio, visual and joint modalities. We conduct experiments on AVS benchmarks, demonstrating that our method achieves state-of-the-art (SOTA) segmentation accuracy.
Authors: Zhipeng Qian, Yiwei Ma, Zhekai Lin, Jiayi Ji, Xiawu Zheng, Xiaoshuai Sun, Rongrong Ji
Abstract: 3D referring expression comprehension (3DREC) and segmentation (3DRES) have overlapping objectives, indicating their potential for collaboration. However, existing collaborative approaches predominantly depend on the results of one task to make predictions for the other, limiting effective collaboration. We argue that employing separate branches for 3DREC and 3DRES tasks enhances the model's capacity to learn specific information for each task, enabling them to acquire complementary knowledge. Thus, we propose the MCLN framework, which includes independent branches for 3DREC and 3DRES tasks. This enables dedicated exploration of each task and effective coordination between the branches. Furthermore, to facilitate mutual reinforcement between these branches, we introduce a Relative Superpoint Aggregation (RSA) module and an Adaptive Soft Alignment (ASA) module. These modules significantly contribute to the precise alignment of prediction results from the two branches, directing the module to allocate increased attention to key positions. Comprehensive experimental evaluation demonstrates that our proposed method achieves state-of-the-art performance on both the 3DREC and 3DRES tasks, with an increase of 3.27% in Acc@0.5 for 3DREC and 5.22% in mIOU for 3DRES.
Authors: Zhonghang Liu, Panzhong Lu, Guoyang Xie, Zhichao Lu, Wen-Yan Lin
Abstract: In the realm of unsupervised image outlier detection, assigning outlier scores holds greater significance than its subsequent task: thresholding for predicting labels. This is because determining the optimal threshold on non-separable outlier score functions is an ill-posed problem. However, the lack of predicted labels not only hiders some real applications of current outlier detectors but also causes these methods not to be enhanced by leveraging the dataset's self-supervision. To advance existing scoring methods, we propose a multiple thresholding (Multi-T) module. It generates two thresholds that isolate inliers and outliers from the unlabelled target dataset, whereas outliers are employed to obtain better feature representation while inliers provide an uncontaminated manifold. Extensive experiments verify that Multi-T can significantly improve proposed outlier scoring methods. Moreover, Multi-T contributes to a naive distance-based method being state-of-the-art.
Authors: You Wu, Xucheng Wang, Dan Zeng, Hengzhou Ye, Xiaolan Xie, Qijun Zhao, Shuiwang Li
Abstract: Recently, the surge in the adoption of single-stream architectures utilizing pre-trained ViT backbones represents a promising advancement in the field of generic visual tracking. By integrating feature extraction and fusion into a cohesive framework, these architectures offer improved performance, efficiency, and robustness. However, there has been limited exploration into optimizing these frameworks for UAV tracking. In this paper, we boost the efficiency of this framework by tailoring it into an adaptive computation framework that dynamically exits Transformer blocks for real-time UAV tracking. The motivation behind this is that tracking tasks with fewer challenges can be adequately addressed using low-level feature representations. Simpler tasks can often be handled with less demanding, lower-level features. This approach allows the model use computational resources more efficiently by focusing on complex tasks and conserving resources for easier ones. Another significant enhancement introduced in this paper is the improved effectiveness of ViTs in handling motion blur, a common issue in UAV tracking caused by the fast movements of either the UAV, the tracked objects, or both. This is achieved by acquiring motion blur robust representations through enforcing invariance in the feature representation of the target with respect to simulated motion blur. The proposed approach is dubbed BDTrack. Extensive experiments conducted on five tracking benchmarks validate the effectiveness and versatility of our approach, establishing it as a cutting-edge solution in real-time UAV tracking. Code is released at: https://github.com/wuyou3474/BDTrack.
Authors: Qi Sun, Hang Zhou, Wengang Zhou, Li Li, Houqiang Li
Abstract: Synthesizing realistic 3D indoor scenes is a challenging task that traditionally relies on manual arrangement and annotation by expert designers. Recent advances in autoregressive models have automated this process, but they often lack semantic understanding of the relationships and hierarchies present in real-world scenes, yielding limited performance. In this paper, we propose Forest2Seq, a framework that formulates indoor scene synthesis as an order-aware sequential learning problem. Forest2Seq organizes the inherently unordered collection of scene objects into structured, ordered hierarchical scene trees and forests. By employing a clustering-based algorithm and a breadth-first traversal, Forest2Seq derives meaningful orderings and utilizes a transformer to generate realistic 3D scenes autoregressively. Experimental results on standard benchmarks demonstrate Forest2Seq's superiority in synthesizing more realistic scenes compared to top-performing baselines, with significant improvements in FID and KL scores. Our additional experiments for downstream tasks and ablation studies also confirm the importance of incorporating order as a prior in 3D scene generation.
Authors: Xingyang Nie, Su Pan, Xiaoyu Zhai, Shifei Tao, Fengzhong Qu, Biao Wang, Huilin Ge, Guojie Xiao
Abstract: Underwater image enhancement (UIE) has attracted much attention owing to its importance for underwater operation and marine engineering. Motivated by the recent advance in generative models, we propose a novel UIE method based on image-conditional diffusion transformer (ICDT). Our method takes the degraded underwater image as the conditional input and converts it into latent space where ICDT is applied. ICDT replaces the conventional U-Net backbone in a denoising diffusion probabilistic model (DDPM) with a transformer, and thus inherits favorable properties such as scalability from transformers. Furthermore, we train ICDT with a hybrid loss function involving variances to achieve better log-likelihoods, which meanwhile significantly accelerates the sampling process. We experimentally assess the scalability of ICDTs and compare with prior works in UIE on the Underwater ImageNet dataset. Besides good scaling properties, our largest model, ICDT-XL/2, outperforms all comparison methods, achieving state-of-the-art (SOTA) quality of image enhancement.
Authors: Juzheng Miao, Cheng Chen, Keli Zhang, Jie Chuai, Quanzheng Li, Pheng-Ann Heng
Abstract: One-shot detection of anatomical landmarks is gaining significant attention for its efficiency in using minimal labeled data to produce promising results. However, the success of current methods heavily relies on the employment of extensive unlabeled data to pre-train an effective feature extractor, which limits their applicability in scenarios where a substantial amount of unlabeled data is unavailable. In this paper, we propose the first foundation model-enabled one-shot landmark detection (FM-OSD) framework for accurate landmark detection in medical images by utilizing solely a single template image without any additional unlabeled data. Specifically, we use the frozen image encoder of visual foundation models as the feature extractor, and introduce dual-branch global and local feature decoders to increase the resolution of extracted features in a coarse to fine manner. The introduced feature decoders are efficiently trained with a distance-aware similarity learning loss to incorporate domain knowledge from the single template image. Moreover, a novel bidirectional matching strategy is developed to improve both robustness and accuracy of landmark detection in the case of scattered similarity map obtained by foundation models. We validate our method on two public anatomical landmark detection datasets. By using solely a single template image, our method demonstrates significant superiority over strong state-of-the-art one-shot landmark detection methods.
Authors: Yiran Ling, Rongqiang Zhao, Yixuan Shen, Dongbo Li, Jing Jin, Jie Liu
Abstract: Non-contact volume estimation of pile-type objects has considerable potential in industrial scenarios, including grain, coal, mining, and stone materials. However, using existing method for these scenarios is challenged by unstable measurement poses, significant light interference, the difficulty of training data collection, and the computational burden brought by large piles. To address the above issues, we propose the Depth Integrated Volume EStimation of Pile Of Things (DIVESPOT) based on point cloud technology in this study. For the challenges of unstable measurement poses, the point cloud pose correction and filtering algorithm is designed based on the Random Sample Consensus (RANSAC) and the Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN). To cope with light interference and to avoid the relying on training data, the height-distribution-based ground feature extraction algorithm is proposed to achieve RGB-independent. To reduce the computational burden, the storage space optimizing strategy is developed, such that accurate estimation can be acquired by using compressed voxels. Experimental results demonstrate that the DIVESPOT method enables non-data-driven, RGB-independent segmentation of pile point clouds, maintaining a volume calculation relative error within 2%. Even with 90% compression of the voxel mesh, the average error of the results can be under 3%.
Authors: Juzheng Miao, Cheng Chen, Keli Zhang, Jie Chuai, Quanzheng Li, Pheng-Ann Heng
Abstract: Semi-supervised learning (SSL) has achieved notable progress in medical image segmentation. To achieve effective SSL, a model needs to be able to efficiently learn from limited labeled data and effectively exploiting knowledge from abundant unlabeled data. Recent developments in visual foundation models, such as the Segment Anything Model (SAM), have demonstrated remarkable adaptability with improved sample efficiency. To harness the power of foundation models for application in SSL, we propose a cross prompting consistency method with segment anything model (CPC-SAM) for semi-supervised medical image segmentation. Our method employs SAM's unique prompt design and innovates a cross-prompting strategy within a dual-branch framework to automatically generate prompts and supervisions across two decoder branches, enabling effectively learning from both scarce labeled and valuable unlabeled data. We further design a novel prompt consistency regularization, to reduce the prompt position sensitivity and to enhance the output invariance under different prompts. We validate our method on two medical image segmentation tasks. The extensive experiments with different labeled-data ratios and modalities demonstrate the superiority of our proposed method over the state-of-the-art SSL methods, with more than 9% Dice improvement on the breast cancer segmentation task.
Authors: Keke Zu, Hu Zhang, Jian Lu, Lei Zhang, Chen Xu
Abstract: This work presents a novel module, namely multi-branch concat (MBC), to process the input tensor and obtain the multi-scale feature map. The proposed MBC module brings new degrees of freedom (DoF) for the design of attention networks by allowing the type of transformation operators and the number of branches to be flexibly adjusted. Two important transformation operators, multiplex and split, are considered in this work, both of which can represent multi-scale features at a more granular level and increase the range of receptive fields. By integrating the MBC and attention module, a multi-branch attention (MBA) module is consequently developed to capture the channel-wise interaction of feature maps for establishing the long-range channel dependency. By substituting the 3x3 convolutions in the bottleneck blocks of the ResNet with the proposed MBA, a novel block namely efficient multi-branch attention (EMBA) is obtained, which can be easily plugged into the state-of-the-art backbone CNN models. Furthermore, a new backbone network called EMBANet is established by stacking the EMBA blocks. The proposed EMBANet is extensively evaluated on representative computer vision tasks including: classification, detection, and segmentation. And it demonstrates consistently superior performance over the popular backbones.
Authors: Nafisa Hussain
Abstract: Videos serve as a powerful medium to convey ideas, tell stories, and provide detailed instructions, especially through long-format tutorials. Such tutorials are valuable for learning new skills at one's own pace, yet they can be overwhelming due to their length and dense content. Viewers often seek specific information, like precise measurements or step-by-step execution details, making it essential to extract and summarize key segments efficiently. An intelligent, time-sensitive video assistant capable of summarizing and detecting highlights in long videos is highly sought after. Recent advancements in Multimodal Large Language Models offer promising solutions to develop such an assistant. Our research explores the use of multimodal models to enhance video summarization and step-by-step instruction generation within specific domains. These models need to understand temporal events and relationships among actions across video frames. Our approach focuses on fine-tuning TimeChat to improve its performance in specific domains: cooking and medical procedures. By training the model on domain-specific datasets like Tasty for cooking and MedVidQA for medical procedures, we aim to enhance its ability to generate concise, accurate summaries of instructional videos. We curate and restructure these datasets to create high-quality video-centric instruction data. Our findings indicate that when finetuned on domain-specific procedural data, TimeChat can significantly improve the extraction and summarization of key instructional steps in long-format videos. This research demonstrates the potential of specialized multimodal models to assist with practical tasks by providing personalized, step-by-step guidance tailored to the unique aspects of each domain.
Authors: Idris Hamoud, Alexandros Karargyris, Aidean Sharghi, Omid Mohareri, Nicolas Padoy
Abstract: Semantic segmentation and activity classification are key components to creating intelligent surgical systems able to understand and assist clinical workflow. In the Operating Room, semantic segmentation is at the core of creating robots aware of clinical surroundings, whereas activity classification aims at understanding OR workflow at a higher level. State-of-the-art semantic segmentation and activity recognition approaches are fully supervised, which is not scalable. Self-supervision can decrease the amount of annotated data needed. We propose a new 3D self-supervised task for OR scene understanding utilizing OR scene images captured with ToF cameras. Contrary to other self-supervised approaches, where handcrafted pretext tasks are focused on 2D image features, our proposed task consists of predicting the relative 3D distance of image patches by exploiting the depth maps. Learning 3D spatial context generates discriminative features for our downstream tasks. Our approach is evaluated on two tasks and datasets containing multi-view data captured from clinical scenarios. We demonstrate a noteworthy improvement of performance on both tasks, specifically on low-regime data where utility of self-supervised learning is the highest.
Authors: Tuan T. Nguyen, Phan Le, Yasir Hassan, Mina Sartipi
Abstract: In this paper, we present the submission to the 5th Annual Smoky Mountains Computational Sciences Data Challenge, Challenge 3. This is the solution for semantic segmentation problem in both real-world and synthetic images from a vehicle s forward-facing camera. We concentrate in building a robust model which performs well across various domains of different outdoor situations such as sunny, snowy, rainy, etc. In particular, our method is developed with two main directions: model development and domain adaptation. In model development, we use the High Resolution Network (HRNet) as the baseline. Then, this baseline s result is processed by two coarse-to-fine models: Object-Contextual Representations (OCR) and Hierarchical Multi-scale Attention (HMA) to get the better robust feature. For domain adaption, we implement the Domain-Based Batch Normalization (DNB) to reduce the distribution shift from diverse domains. Our proposed method yield 81.259 mean intersection-over-union (mIoU) in validation set. This paper studies the effectiveness of employing real-world and synthetic data to handle the domain adaptation in semantic segmentation problem.
Authors: Tuan T. Nguyen, Mina Sartipi
Abstract: Given the rising urban population and the consequential rise in traffic congestion, the implementation of smart parking systems has emerged as a critical matter of concern. Smart parking solutions use cameras, sensors, and algorithms like computer vision to find available parking spaces. This method improves parking place recognition, reduces traffic and pollution, and optimizes travel time. In recent years, computer vision-based approaches have been widely used. However, most existing studies rely on manually labeled parking spots, which has implications for the cost and practicality of implementation. To solve this problem, we propose a novel approach PakLoc, which automatically localize parking spots. Furthermore, we present the PakSke module, which automatically adjust the rotation and the size of detected bounding box. The efficacy of our proposed methodology on the PKLot dataset results in a significant reduction in human labor of 94.25\%. Another fundamental aspect of a smart parking system is its capacity to accurately determine and indicate the state of parking spots within a parking lot. The conventional approach involves employing classification techniques to forecast the condition of parking spots based on the bounding boxes derived from manually labeled grids. In this study, we provide a novel approach called PakSta for identifying the state of parking spots automatically. Our method utilizes object detector from PakLoc to simultaneously determine the occupancy status of all parking lots within a video frame. Our proposed method PakSta exhibits a competitive performance on the PKLot dataset when compared to other classification methods.
Authors: Athena Psalta, Vasileios Tsironis, Andreas El Saer, Konstantinos Karantzalos
Abstract: Object tracking in satellite videos remains a complex endeavor in remote sensing due to the intricate and dynamic nature of satellite imagery. Existing state-of-the-art trackers in computer vision integrate sophisticated architectures, attention mechanisms, and multi-modal fusion to enhance tracking accuracy across diverse environments. However, the challenges posed by satellite imagery, such as background variations, atmospheric disturbances, and low-resolution object delineation, significantly impede the precision and reliability of traditional Single Object Tracking (SOT) techniques. Our study delves into these challenges and proposes prompt engineering methodologies, leveraging the Segment Anything Model (SAM) and TAPIR (Tracking Any Point with per-frame Initialization and temporal Refinement), to create a training-free point-based tracking method for small-scale objects on satellite videos. Experiments on the VISO dataset validate our strategy, marking a significant advancement in robust tracking solutions tailored for satellite imagery in remote sensing applications.
Authors: Seung Park, Yong-Goo Shin
Abstract: Various models based on StyleGAN have gained significant traction in the field of image synthesis, attributed to their robust training stability and superior performances. Within the StyleGAN framework, the adoption of image skip connection is favored over the traditional residual connection. However, this preference is just based on empirical observations; there has not been any in-depth mathematical analysis on it yet. To rectify this situation, this brief aims to elucidate the mathematical meaning of the image skip connection and introduce a groundbreaking methodology, termed the image squeeze connection, which significantly improves the quality of image synthesis. Specifically, we analyze the image skip connection technique to reveal its problem and introduce the proposed method which not only effectively boosts the GAN performance but also reduces the required number of network parameters. Extensive experiments on various datasets demonstrate that the proposed method consistently enhances the performance of state-of-the-art models based on StyleGAN. We believe that our findings represent a vital advancement in the field of image synthesis, suggesting a novel direction for future research and applications.
Authors: Paul Albert, Jack Valmadre, Eric Arazo, Tarun Krishna, Noel E. O'Connor, Kevin McGuinness
Abstract: Training a classifier on web-crawled data demands learning algorithms that are robust to annotation errors and irrelevant examples. This paper builds upon the recent empirical observation that applying unsupervised contrastive learning to noisy, web-crawled datasets yields a feature representation under which the in-distribution (ID) and out-of-distribution (OOD) samples are linearly separable. We show that direct estimation of the separating hyperplane can indeed offer an accurate detection of OOD samples, and yet, surprisingly, this detection does not translate into gains in classification accuracy. Digging deeper into this phenomenon, we discover that the near-perfect detection misses a type of clean examples that are valuable for supervised learning. These examples often represent visually simple images, which are relatively easy to identify as clean examples using standard loss- or distance-based methods despite being poorly separated from the OOD distribution using unsupervised learning. Because we further observe a low correlation with SOTA metrics, this urges us to propose a hybrid solution that alternates between noise detection using linear separation and a state-of-the-art (SOTA) small-loss approach. When combined with the SOTA algorithm PLS, we substantially improve SOTA results for real-world image classification in the presence of web noise github.com/PaulAlbert31/LSA
Authors: Chenxin Li, Xinyu Liu, Cheng Wang, Yifan Liu, Weihao Yu, Jing Shao, Yixuan Yuan
Abstract: Recent advances in learning multi-modal representation have witnessed the success in biomedical domains. While established techniques enable handling multi-modal information, the challenges are posed when extended to various clinical modalities and practical modalitymissing setting due to the inherent modality gaps. To tackle these, we propose an innovative Modality-prompted Heterogeneous Graph for Omnimodal Learning (GTP-4o), which embeds the numerous disparate clinical modalities into a unified representation, completes the deficient embedding of missing modality and reformulates the cross-modal learning with a graph-based aggregation. Specially, we establish a heterogeneous graph embedding to explicitly capture the diverse semantic properties on both the modality-specific features (nodes) and the cross-modal relations (edges). Then, we design a modality-prompted completion that enables completing the inadequate graph representation of missing modality through a graph prompting mechanism, which generates hallucination graphic topologies to steer the missing embedding towards the intact representation. Through the completed graph, we meticulously develop a knowledge-guided hierarchical cross-modal aggregation consisting of a global meta-path neighbouring to uncover the potential heterogeneous neighbors along the pathways driven by domain knowledge, and a local multi-relation aggregation module for the comprehensive cross-modal interaction across various heterogeneous relations. We assess the efficacy of our methodology on rigorous benchmarking experiments against prior state-of-the-arts. In a nutshell, GTP-4o presents an initial foray into the intriguing realm of embedding, relating and perceiving the heterogeneous patterns from various clinical modalities holistically via a graph theory. Project page: https://gtp-4-o.github.io/.
Authors: Sherry X. Chen, Yaron Vaxman, Elad Ben Baruch, David Asulin, Aviad Moreshet, Misha Sra, Pradeep Sen
Abstract: We propose Image Content Appeal Assessment (ICAA), a novel metric that quantifies the level of positive interest an image's content generates for viewers, such as the appeal of food in a photograph. This is fundamentally different from traditional Image-Aesthetics Assessment (IAA), which judges an image's artistic quality. While previous studies often confuse the concepts of ``aesthetics'' and ``appeal,'' our work addresses this by being the first to study ICAA explicitly. To do this, we propose a novel system that automates dataset creation and implements algorithms to estimate and boost content appeal. We use our pipeline to generate two large-scale datasets (70K+ images each) in diverse domains (food and room interior design) to train our models, which revealed little correlation between content appeal and aesthetics. Our user study, with more than 76% of participants preferring the appeal-enhanced images, confirms that our appeal ratings accurately reflect user preferences, establishing ICAA as a unique evaluative criterion. Our code and datasets are available at https://github.com/SherryXTChen/AID-Appeal.
Authors: Kanghao Chen, Hangyu Li, JiaZhou Zhou, Zeyu Wang, Lin Wang
Abstract: Event cameras harness advantages such as low latency, high temporal resolution, and high dynamic range (HDR), compared to standard cameras. Due to the distinct imaging paradigm shift, a dominant line of research focuses on event-to-video (E2V) reconstruction to bridge event-based and standard computer vision. However, this task remains challenging due to its inherently ill-posed nature: event cameras only detect the edge and motion information locally. Consequently, the reconstructed videos are often plagued by artifacts and regional blur, primarily caused by the ambiguous semantics of event data. In this paper, we find language naturally conveys abundant semantic information, rendering it stunningly superior in ensuring semantic consistency for E2V reconstruction. Accordingly, we propose a novel framework, called LaSe-E2V, that can achieve semantic-aware high-quality E2V reconstruction from a language-guided perspective, buttressed by the text-conditional diffusion models. However, due to diffusion models' inherent diversity and randomness, it is hardly possible to directly apply them to achieve spatial and temporal consistency for E2V reconstruction. Thus, we first propose an Event-guided Spatiotemporal Attention (ESA) module to condition the event data to the denoising pipeline effectively. We then introduce an event-aware mask loss to ensure temporal coherence and a noise initialization strategy to enhance spatial consistency. Given the absence of event-text-video paired data, we aggregate existing E2V datasets and generate textual descriptions using the tagging models for training and evaluation. Extensive experiments on three datasets covering diverse challenging scenarios (e.g., fast motion, low light) demonstrate the superiority of our method. Dataset and code will be available upon acceptance.
Authors: Yujin Jeong, Yunji Kim, Sanghyuk Chun, Jiyoung Lee
Abstract: Multimodal generative models have shown impressive advances with the help of powerful diffusion models. Despite the progress, generating sound solely from text poses challenges in ensuring comprehensive scene depiction and temporal alignment. Meanwhile, video-to-sound generation limits the flexibility to prioritize sound synthesis for specific objects within the scene. To tackle these challenges, we propose a novel video-and-text-to-sound generation method, called ReWaS, where video serves as a conditional control for a text-to-audio generation model. Our method estimates the structural information of audio (namely, energy) from the video while receiving key content cues from a user prompt. We employ a well-performing text-to-sound model to consolidate the video control, which is much more efficient for training multimodal diffusion models with massive triplet-paired (audio-video-text) data. In addition, by separating the generative components of audio, it becomes a more flexible system that allows users to freely adjust the energy, surrounding environment, and primary sound source according to their preferences. Experimental results demonstrate that our method shows superiority in terms of quality, controllability, and training efficiency. Our demo is available at https://naver-ai.github.io/rewas
Authors: Jia Liu, Changlin Li, Qirui Sun, Jiahui Ming, Chen Fang, Jue Wang, Bing Zeng, Shuaicheng Liu
Abstract: Fine-tuning advanced diffusion models for high-quality image stylization usually requires large training datasets and substantial computational resources, hindering their practical applicability. We propose Ada-Adapter, a novel framework for few-shot style personalization of diffusion models. Ada-Adapter leverages off-the-shelf diffusion models and pre-trained image feature encoders to learn a compact style representation from a limited set of source images. Our method enables efficient zero-shot style transfer utilizing a single reference image. Furthermore, with a small number of source images (three to five are sufficient) and a few minutes of fine-tuning, our method can capture intricate style details and conceptual characteristics, generating high-fidelity stylized images that align well with the provided text prompts. We demonstrate the effectiveness of our approach on various artistic styles, including flat art, 3D rendering, and logo design. Our experimental results show that Ada-Adapter outperforms existing zero-shot and few-shot stylization methods in terms of output quality, diversity, and training efficiency.
Authors: Yafei Mao, Christopher Merkle, Jan P. Allebach
Abstract: This paper presents an approach to predict the color of skin-with-foundation based on a no makeup selfie image and a foundation shade image. Our approach first calibrates the image with the help of the color checker target, and then trains a supervised-learning model to predict the skin color. In the calibration stage, We propose to use three different transformation matrices to map the device dependent RGB response to the reference CIE XYZ space. In so doing, color correction error can be minimized. We then compute the average value of the region of interest in the calibrated images, and feed them to the prediction model. We explored both the linear regression and support vector regression models. Cross-validation results show that both models can accurately make the prediction.
Authors: Qingyao Tian, Zhen Chen, Huai Liao, Xinyan Huang, Bingyu Yang, Lujie Li, Hongbin Liu
Abstract: Accurate bronchoscope localization is essential for pulmonary interventions, by providing six degrees of freedom (DOF) in airway navigation. However, the robustness of current vision-based methods is often compromised in clinical practice, and they struggle to perform in real-time and to generalize across cases unseen during training. To overcome these challenges, we propose a novel Probabilistic Airway Navigation System (PANS), leveraging Monte-Carlo method with pose hypotheses and likelihoods to achieve robust and real-time bronchoscope localization. Specifically, our PANS incorporates diverse visual representations (\textit{e.g.}, odometry and landmarks) by leveraging two key modules, including the Depth-based Motion Inference (DMI) and the Bronchial Semantic Analysis (BSA). To generate the pose hypotheses of bronchoscope for PANS, we devise the DMI to accurately propagate the estimation of pose hypotheses over time. Moreover, to estimate the accurate pose likelihood, we devise the BSA module by effectively distinguishing between similar bronchial regions in endoscopic images, along with a novel metric to assess the congruence between estimated depth maps and the segmented airway structure. Under this probabilistic formulation, our PANS is capable of achieving the 6-DOF bronchoscope localization with superior accuracy and robustness. Extensive experiments on the collected pulmonary intervention dataset comprising 10 clinical cases confirm the advantage of our PANS over state-of-the-arts, in terms of both robustness and generalization in localizing deeper airway branches and the efficiency of real-time inference. The proposed PANS reveals its potential to be a reliable tool in the operating room, promising to enhance the quality and safety of pulmonary interventions.
Authors: Bangbang Zhou, Yadong Qu, Zixiao Wang, Zicheng Li, Boqiang Zhang, Hongtao Xie
Abstract: Recently, scene text recognition (STR) models have shown significant performance improvements. However, existing models still encounter difficulties in recognizing challenging texts that involve factors such as severely distorted and perspective characters. These challenging texts mainly cause two problems: (1) Large Intra-Class Variance. (2) Small Inter-Class Variance. An extremely distorted character may prominently differ visually from other characters within the same category, while the variance between characters from different classes is relatively small. To address the above issues, we propose a novel method that enriches the character features to enhance the discriminability of characters. Firstly, we propose the Character-Aware Constraint Encoder (CACE) with multiple blocks stacked. CACE introduces a decay matrix in each block to explicitly guide the attention region for each token. By continuously employing the decay matrix, CACE enables tokens to perceive morphological information at the character level. Secondly, an Intra-Inter Consistency Loss (I^2CL) is introduced to consider intra-class compactness and inter-class separability at feature space. I^2CL improves the discriminative capability of features by learning a long-term memory unit for each character category. Trained with synthetic data, our model achieves state-of-the-art performance on common benchmarks (94.1% accuracy) and Union14M-Benchmark (61.6% accuracy). Code is available at https://github.com/bang123-box/CFE.
Authors: Xuan Wang, Hao Tang, Zhigang Zhu
Abstract: Various contextual information has been employed by many approaches for visual detection tasks. However, most of the existing approaches only focus on specific context for specific tasks. In this paper, GMC, a general framework is proposed for multistage context learning and utilization, with various deep network architectures for various visual detection tasks. The GMC framework encompasses three stages: preprocessing, training, and post-processing. In the preprocessing stage, the representation of local context is enhanced by utilizing commonly used labeling standards. During the training stage, semantic context information is fused with visual information, leveraging prior knowledge from the training dataset to capture semantic relationships. In the post-processing stage, general topological relations and semantic masks for stuff are incorporated to enable spatial context reasoning between objects. The proposed framework provides a comprehensive and adaptable solution for context learning and utilization in visual detection scenarios. The framework offers flexibility with user-defined configurations and provide adaptability to diverse network architectures and visual detection tasks, offering an automated and streamlined solution that minimizes user effort and inference time in context learning and reasoning. Experimental results on the visual detection tasks, for storefront object detection, pedestrian detection and COCO object detection, demonstrate that our framework outperforms previous state-of-the-art detectors and transformer architectures. The experiments also demonstrate that three contextual learning components can not only be applied individually and in combination, but can also be applied to various network architectures, and its flexibility and effectiveness in various detection scenarios.
Authors: Tingyu Liu, Jun Huang, Chenyi Weng
Abstract: Inferring future activity information based on observed activity data is a crucial step to improve the accuracy of early activity prediction. Traditional methods based on generative adversarial networks(GAN) or joint learning frameworks can achieve good prediction accuracy under low observation ratios, but they usually have high computational costs. In view of this, this paper proposes a spatio-temporal encoding and decoding-based method for future human activity skeleton synthesis. Firstly, algorithms such as time control, discrete cosine transform, and low-pass filtering are used to cut or pad the skeleton sequences. Secondly, the encoder and decoder are responsible for extracting intermediate semantic encoding from observed skeleton sequences and inferring future sequences from the intermediate semantic encoding, respectively. Finally, joint displacement error, velocity error, and acceleration error, three higher-order kinematic features, are used as key components of the loss function to optimize model parameters. Experimental results show that the proposed future skeleton synthesis algorithm performs better than some existing algorithms. It generates skeleton sequences with smaller errors and fewer model parameters, effectively providing future information for early activity prediction.
Authors: Zhongtian Wang, You Wu, Hui Zhou, Shuiwang Li
Abstract: Object detection has greatly improved over the past decade thanks to advances in deep learning and large-scale datasets. However, detecting objects reflected in surfaces remains an underexplored area. Reflective surfaces are ubiquitous in daily life, appearing in homes, offices, public spaces, and natural environments. Accurate detection and interpretation of reflected objects are essential for various applications. This paper addresses this gap by introducing a extensive benchmark specifically designed for Reflected Object Detection. Our Reflected Object Detection Dataset (RODD) features a diverse collection of images showcasing reflected objects in various contexts, providing standard annotations for both real and reflected objects. This distinguishes it from traditional object detection benchmarks. RODD encompasses 10 categories and includes 21,059 images of real and reflected objects across different backgrounds, complete with standard bounding box annotations and the classification of objects as real or reflected. Additionally, we present baseline results by adapting five state-of-the-art object detection models to address this challenging task. Experimental results underscore the limitations of existing methods when applied to reflected object detection, highlighting the need for specialized approaches. By releasing RODD, we aim to support and advance future research on detecting reflected objects. Dataset and code are available at: https: //github.com/Tqybu-hans/RODD.
Authors: Yuejiao Su, Yi Wang, Lap-Pui Chau
Abstract: Egocentric hand-object segmentation (EgoHOS) is a brand-new task aiming at segmenting the hands and interacting objects in the egocentric image. Although significant advancements have been achieved by current methods, establishing an end-to-end model with high accuracy remains an unresolved challenge. Moreover, existing methods lack explicit modeling of the relationships between hands and objects as well as objects and objects, thereby disregarding critical information on hand-object interaction and introducing confusion into algorithms, ultimately leading to a reduction in segmentation performance. To address the limitations of existing methods, this paper proposes a novel end-to-end Object-centric Relationship Modeling Network (ORMNet) for EgoHOS. Specifically, based on a single-encoder and multi-decoder framework, we design the Hand-Object Relation (HOR) module to leverage hand-guided attention to capture the correlation between hands and objects and facilitate their representations. Moreover, based on the observed interrelationships between diverse categories of objects, we introduce the Object Relation Decoupling (ORD) strategy. This strategy allows the decoupling of the two-hand object during training, thereby alleviating the ambiguity of the network. Experimental results on three datasets show that the proposed ORMNet has notably exceptional segmentation performance with robust generalization capabilities.
Authors: Jiacheng Su, Kunhong Liu, Liyan Chen, Junfeng Yao, Qingsong Liu, Dongdong Lv
Abstract: The existing methods for audio-driven talking head video editing have the limitations of poor visual effects. This paper tries to tackle this problem through editing talking face images seamless with different emotions based on two modules: (1) an audio-to-landmark module, consisting of the CrossReconstructed Emotion Disentanglement and an alignment network module. It bridges the gap between speech and facial motions by predicting corresponding emotional landmarks from speech; (2) a landmark-based editing module edits face videos via StyleGAN. It aims to generate the seamless edited video consisting of the emotion and content components from the input audio. Extensive experiments confirm that compared with state-of-the-arts methods, our method provides high-resolution videos with high visual quality.
Authors: Jiedong Zhuang, Jiaqi Hu, Lianrui Mu, Rui Hu, Xiaoyu Liang, Jiangnan Ye, Haoji Hu
Abstract: CLIP has achieved impressive zero-shot performance after pre-training on a large-scale dataset consisting of paired image-text data. Previous works have utilized CLIP by incorporating manually designed visual prompts like colored circles and blur masks into the images to guide the model's attention, showing enhanced zero-shot performance in downstream tasks. Although these methods have achieved promising results, they inevitably alter the original information of the images, which can lead to failure in specific tasks. We propose a train-free method Foveal-Attention CLIP (FALIP), which adjusts the CLIP's attention by inserting foveal attention masks into the multi-head self-attention module. We demonstrate FALIP effectively boosts CLIP zero-shot performance in tasks such as referring expressions comprehension, image classification, and 3D point cloud recognition. Experimental results further show that FALIP outperforms existing methods on most metrics and can augment current methods to enhance their performance.
Authors: Xianrui Luo, Huiqiang Sun, Juewen Peng, Zhiguo Cao
Abstract: Dynamic Neural Radiance Field (NeRF) from monocular videos has recently been explored for space-time novel view synthesis and achieved excellent results. However, defocus blur caused by depth variation often occurs in video capture, compromising the quality of dynamic reconstruction because the lack of sharp details interferes with modeling temporal consistency between input views. To tackle this issue, we propose D2RF, the first dynamic NeRF method designed to restore sharp novel views from defocused monocular videos. We introduce layered Depth-of-Field (DoF) volume rendering to model the defocus blur and reconstruct a sharp NeRF supervised by defocused views. The blur model is inspired by the connection between DoF rendering and volume rendering. The opacity in volume rendering aligns with the layer visibility in DoF rendering.To execute the blurring, we modify the layered blur kernel to the ray-based kernel and employ an optimized sparse kernel to gather the input rays efficiently and render the optimized rays with our layered DoF volume rendering. We synthesize a dataset with defocused dynamic scenes for our task, and extensive experiments on our dataset show that our method outperforms existing approaches in synthesizing all-in-focus novel views from defocus blur while maintaining spatial-temporal consistency in the scene.
Authors: Zehui Zhao, Laith Alzubaidi, Jinglan Zhang, Ye Duan, Usman Naseem, Yuantong Gu
Abstract: Recently, transfer learning and self-supervised learning have gained significant attention within the medical field due to their ability to mitigate the challenges posed by limited data availability, improve model generalisation, and reduce computational expenses. Transfer learning and self-supervised learning hold immense potential for advancing medical research. However, it is crucial to recognise that transfer learning and self-supervised learning architectures exhibit distinct advantages and limitations, manifesting variations in accuracy, training speed, and robustness. This paper compares the performance and robustness of transfer learning and self-supervised learning in the medical field. Specifically, we pre-trained two models using the same source domain datasets with different pre-training methods and evaluated them on small-sized medical datasets to identify the factors influencing their final performance. We tested data with several common issues in medical domains, such as data imbalance, data scarcity, and domain mismatch, through comparison experiments to understand their impact on specific pre-trained models. Finally, we provide recommendations to help users apply transfer learning and self-supervised learning methods in medical areas, and build more convenient and efficient deployment strategies.
Authors: Xiwei Xuan, Ziquan Deng, Hsuan-Tien Lin, Kwan-Liu Ma
Abstract: Recent studies highlight that deep learning models often learn spurious features mistakenly linked to labels, compromising their reliability in real-world scenarios where such correlations do not hold. Despite the increasing research effort, existing solutions often face two main challenges: they either demand substantial annotations of spurious attributes, or they yield less competitive outcomes with expensive training when additional annotations are absent. In this paper, we introduce SLIM, a cost-effective and performance-targeted approach to reducing spurious correlations in deep learning. Our method leverages a human-in-the-loop protocol featuring a novel attention labeling mechanism with a constructed attention representation space. SLIM significantly reduces the need for exhaustive additional labeling, requiring human input for fewer than 3% of instances. By prioritizing data quality over complicated training strategies, SLIM curates a smaller yet more feature-balanced data subset, fostering the development of spuriousness-robust models. Experimental validations across key benchmarks demonstrate that SLIM competes with or exceeds the performance of leading methods while significantly reducing costs. The SLIM framework thus presents a promising path for developing reliable models more efficiently. Our code is available in https://github.com/xiweix/SLIM.git/.
Authors: Weiyi Xue, Zehan Zheng, Fan Lu, Haiyun Wei, Guang Chen, Changjun Jiang
Abstract: Although recent efforts have extended Neural Radiance Fields (NeRF) into LiDAR point cloud synthesis, the majority of existing works exhibit a strong dependence on precomputed poses. However, point cloud registration methods struggle to achieve precise global pose estimation, whereas previous pose-free NeRFs overlook geometric consistency in global reconstruction. In light of this, we explore the geometric insights of point clouds, which provide explicit registration priors for reconstruction. Based on this, we propose Geometry guided Neural LiDAR Fields(GeoNLF), a hybrid framework performing alternately global neural reconstruction and pure geometric pose optimization. Furthermore, NeRFs tend to overfit individual frames and easily get stuck in local minima under sparse-view inputs. To tackle this issue, we develop a selective-reweighting strategy and introduce geometric constraints for robust optimization. Extensive experiments on NuScenes and KITTI-360 datasets demonstrate the superiority of GeoNLF in both novel view synthesis and multi-view registration of low-frequency large-scale point clouds.
Authors: Zhenyu Wang, Aoxue Li, Zhenguo Li, Xihui Liu
Abstract: Despite the success achieved by existing image generation and editing methods, current models still struggle with complex problems including intricate text prompts, and the absence of verification and self-correction mechanisms makes the generated images unreliable. Meanwhile, a single model tends to specialize in particular tasks and possess the corresponding capabilities, making it inadequate for fulfilling all user requirements. We propose GenArtist, a unified image generation and editing system, coordinated by a multimodal large language model (MLLM) agent. We integrate a comprehensive range of existing models into the tool library and utilize the agent for tool selection and execution. For a complex problem, the MLLM agent decomposes it into simpler sub-problems and constructs a tree structure to systematically plan the procedure of generation, editing, and self-correction with step-by-step verification. By automatically generating missing position-related inputs and incorporating position information, the appropriate tool can be effectively employed to address each sub-problem. Experiments demonstrate that GenArtist can perform various generation and editing tasks, achieving state-of-the-art performance and surpassing existing models such as SDXL and DALL-E 3, as can be seen in Fig. 1. Project page is https://zhenyuw16.github.io/GenArtist_page.
Authors: Pingyi Chen, Chenglu Zhu, Sunyi Zheng, Honglin Li, Lin Yang
Abstract: Whole slide imaging is routinely adopted for carcinoma diagnosis and prognosis. Abundant experience is required for pathologists to achieve accurate and reliable diagnostic results of whole slide images (WSI). The huge size and heterogeneous features of WSIs make the workflow of pathological reading extremely time-consuming. In this paper, we propose a novel framework (WSI-VQA) to interpret WSIs by generative visual question answering. WSI-VQA shows universality by reframing various kinds of slide-level tasks in a question-answering pattern, in which pathologists can achieve immunohistochemical grading, survival prediction, and tumor subtyping following human-machine interaction. Furthermore, we establish a WSI-VQA dataset which contains 8672 slide-level question-answering pairs with 977 WSIs. Besides the ability to deal with different slide-level tasks, our generative model which is named Wsi2Text Transformer (W2T) outperforms existing discriminative models in medical correctness, which reveals the potential of our model to be applied in the clinical scenario. Additionally, we also visualize the co-attention mapping between word embeddings and WSIs as an intuitive explanation for diagnostic results. The dataset and related code are available at https://github.com/cpystan/WSI-VQA.
Authors: Anh-Dzung Doan, Bach Long Nguyen, Terry Lim, Madhuka Jayawardhana, Surabhi Gupta, Christophe Guettier, Ian Reid, Markus Wagner, Tat-Jun Chin
Abstract: Prior to deployment, an object detector is trained on a dataset compiled from a previous data collection campaign. However, the environment in which the object detector is deployed will invariably evolve, particularly in outdoor settings where changes in lighting, weather and seasons will significantly affect the appearance of the scene and target objects. It is almost impossible for all potential scenarios that the object detector may come across to be present in a finite training dataset. This necessitates continuous updates to the object detector to maintain satisfactory performance. Test-time domain adaptation techniques enable machine learning models to self-adapt based on the distributions of the testing data. However, existing methods mainly focus on fully automated adaptation, which makes sense for applications such as self-driving cars. Despite the prevalence of fully automated approaches, in some applications such as surveillance, there is usually a human operator overseeing the system's operation. We propose to involve the operator in test-time domain adaptation to raise the performance of object detection beyond what is achievable by fully automated adaptation. To reduce manual effort, the proposed method only requires the operator to provide weak labels, which are then used to guide the adaptation process. Furthermore, the proposed method can be performed in a streaming setting, where each online sample is observed only once. We show that the proposed method outperforms existing works, demonstrating a great benefit of human-in-the-loop test-time domain adaptation. Our code is publicly available at https://github.com/dzungdoan6/WSTTA
Authors: Wei Ji, Xiangyan Liu, Yingfei Sun, Jiajun Deng, You Qin, Ammar Nuwanna, Mengyao Qiu, Lina Wei, Roger Zimmermann
Abstract: Detecting visual content on language expression has become an emerging topic in the community. However, in the video domain, the existing setting, i.e., spatial-temporal video grounding (STVG), is formulated to only detect one pre-existing object in each frame, ignoring the fact that language descriptions can involve none or multiple entities within a video. In this work, we advance the STVG to a more practical setting called described spatial-temporal video detection (DSTVD) by overcoming the above limitation. To facilitate the exploration of DSTVD, we first introduce a new benchmark, namely DVD-ST. Notably, DVD-ST supports grounding from none to many objects onto the video in response to queries and encompasses a diverse range of over 150 entities, including appearance, actions, locations, and interactions. The extensive breadth and diversity of the DVD-ST dataset make it an exemplary testbed for the investigation of DSTVD. In addition to the new benchmark, we further present two baseline methods for our proposed DSTVD task by extending two representative STVG models, i.e., TubeDETR, and STCAT. These extended models capitalize on tubelet queries to localize and track referred objects across the video sequence. Besides, we adjust the training objectives of these models to optimize spatial and temporal localization accuracy and multi-class classification capabilities. Furthermore, we benchmark the baselines on the introduced DVD-ST dataset and conduct extensive experimental analysis to guide future investigation. Our code and benchmark will be publicly available.
Authors: Ziyang Song, Jinxi Li, Bo Yang
Abstract: It has long been challenging to recover the underlying dynamic 3D scene representations from a monocular RGB video. Existing works formulate this problem into finding a single most plausible solution by adding various constraints such as depth priors and strong geometry constraints, ignoring the fact that there could be infinitely many 3D scene representations corresponding to a single dynamic video. In this paper, we aim to learn all plausible 3D scene configurations that match the input video, instead of just inferring a specific one. To achieve this ambitious goal, we introduce a new framework, called OSN. The key to our approach is a simple yet innovative object scale network together with a joint optimization module to learn an accurate scale range for every dynamic 3D object. This allows us to sample as many faithful 3D scene configurations as possible. Extensive experiments show that our method surpasses all baselines and achieves superior accuracy in dynamic novel view synthesis on multiple synthetic and real-world datasets. Most notably, our method demonstrates a clear advantage in learning fine-grained 3D scene geometry. Our code and data are available at https://github.com/vLAR-group/OSN
Authors: Bowen Wang, Liangzhi Li, Jiahao Zhang, Yuta Nakashima, Hajime Nagahara
Abstract: The imperative to comprehend the behaviors of deep learning models is of utmost importance. In this realm, Explainable Artificial Intelligence (XAI) has emerged as a promising avenue, garnering increasing interest in recent years. Despite this, most existing methods primarily depend on gradients or input perturbation, which often fails to embed explanations directly within the model's decision-making process. Addressing this gap, we introduce ESCOUTER, a visually explainable classifier based on the modified slot attention mechanism. ESCOUTER distinguishes itself by not only delivering high classification accuracy but also offering more transparent insights into the reasoning behind its decisions. It differs from prior approaches in two significant aspects: (a) ESCOUTER incorporates explanations into the final confidence scores for each category, providing a more intuitive interpretation, and (b) it offers positive or negative explanations for all categories, elucidating "why an image belongs to a certain category" or "why it does not." A novel loss function specifically for ESCOUTER is designed to fine-tune the model's behavior, enabling it to toggle between positive and negative explanations. Moreover, an area loss is also designed to adjust the size of the explanatory regions for a more precise explanation. Our method, rigorously tested across various datasets and XAI metrics, outperformed previous state-of-the-art methods, solidifying its effectiveness as an explanatory tool.
Authors: Junhao Su, Changpeng Cai, Feiyu Zhu, Chenghao He, Xiaojie Xu, Dongzhi Guan, Chenyang Si
Abstract: Deep neural networks conventionally employ end-to-end backpropagation for their training process, which lacks biological credibility and triggers a locking dilemma during network parameter updates, leading to significant GPU memory use. Supervised local learning, which segments the network into multiple local blocks updated by independent auxiliary networks. However, these methods cannot replace end-to-end training due to lower accuracy, as gradients only propagate within their local block, creating a lack of information exchange between blocks. To address this issue and establish information transfer across blocks, we propose a Momentum Auxiliary Network (MAN) that establishes a dynamic interaction mechanism. The MAN leverages an exponential moving average (EMA) of the parameters from adjacent local blocks to enhance information flow. This auxiliary network, updated through EMA, helps bridge the informational gap between blocks. Nevertheless, we observe that directly applying EMA parameters has certain limitations due to feature discrepancies among local blocks. To overcome this, we introduce learnable biases, further boosting performance. We have validated our method on four image classification datasets (CIFAR-10, STL-10, SVHN, ImageNet), attaining superior performance and substantial memory savings. Notably, our method can reduce GPU memory usage by more than 45\% on the ImageNet dataset compared to end-to-end training, while achieving higher performance. The Momentum Auxiliary Network thus offers a new perspective for supervised local learning. Our code is available at: \url{https://github.com/JunhaoSu0/MAN}.
Authors: Junhao Su, Chenghao He, Feiyu Zhu, Xiaojie Xu, Dongzhi Guan, Chenyang Si
Abstract: Traditional deep learning relies on end-to-end backpropagation for training, but it suffers from drawbacks such as high memory consumption and not aligning with biological neural networks. Recent advancements have introduced locally supervised learning, which divides networks into modules with isolated gradients and trains them locally. However, this approach can lead to performance lag due to limited interaction between these modules, and the design of auxiliary networks occupies a certain amount of GPU memory. To overcome these limitations, we propose a novel model called HPFF that performs hierarchical locally supervised learning and patch-level feature computation on the auxiliary networks. Hierarchical Locally Supervised Learning (HiLo) enables the network to learn features at different granularity levels along their respective local paths. Specifically, the network is divided into two-level local modules: independent local modules and cascade local modules. The cascade local modules combine two adjacent independent local modules, incorporating both updates within the modules themselves and information exchange between adjacent modules. Patch Feature Fusion (PFF) reduces GPU memory usage by splitting the input features of the auxiliary networks into patches for computation. By averaging these patch-level features, it enhances the network's ability to focus more on those patterns that are prevalent across multiple patches. Furthermore, our method exhibits strong generalization capabilities and can be seamlessly integrated with existing techniques. We conduct experiments on CIFAR-10, STL-10, SVHN, and ImageNet datasets, and the results demonstrate that our proposed HPFF significantly outperforms previous approaches, consistently achieving state-of-the-art performance across different datasets. Our code is available at: \url{https://github.com/Zeudfish/HPFF}.
Authors: Erdong Hu, Longteng Guo, Tongtian Yue, Zijia Zhao, Shuning Xue, Jing Liu
Abstract: In computer vision, Image Difference Captioning (IDC) is crucial for accurately describing variations between closely related images. Traditional IDC methods often rely on specialist models, which restrict their applicability across varied contexts. This paper introduces the OneDiff model, a novel generalist approach that utilizes a robust vision-language model architecture, integrating a siamese image encoder with a Visual Delta Module. This innovative configuration allows for the precise detection and articulation of fine-grained differences between image pairs. OneDiff is trained through a dual-phase strategy, encompassing Coupled Sample Training and multi-task learning across a diverse array of data types, supported by our newly developed DiffCap Dataset. This dataset merges real-world and synthetic data, enhancing the training process and bolstering the model's robustness. Extensive testing on diverse IDC benchmarks, such as Spot-the-Diff, CLEVR-Change, and Birds-to-Words, shows that OneDiff consistently outperforms existing state-of-the-art models in accuracy and adaptability, achieving improvements of up to 85\% CIDEr points in average. By setting a new benchmark in IDC, OneDiff paves the way for more versatile and effective applications in detecting and describing visual differences. The code, models, and data will be made publicly available.
Authors: Jiaying Shi, Xuetong Xue, Shenghui Xu
Abstract: The recent CLIP-based methods have shown promising zero-shot and few-shot performance on image classification tasks. Existing approaches such as CoOp and Tip-Adapter only focus on high-level visual features that are fully aligned with textual features representing the ``Summary" of the image. However, the goal of few-shot learning is to classify unseen images of the same category with few labeled samples. Especially, in contrast to high-level representations, local representations (LRs) at low-level are more consistent between seen and unseen samples. Based on this point, we propose the Meta-Feature Adaption method (MF-Adapter) that combines the complementary strengths of both LRs and high-level semantic representations. Specifically, we introduce the Meta-Feature Unit (MF-Unit), which is a simple yet effective local similarity metric to measure category-consistent local context in an inductive manner. Then we train an MF-Adapter to map image features to MF-Unit for adequately generalizing the intra-class knowledge between unseen images and the support set. Extensive experiments show that our proposed method is superior to the state-of-the-art CLIP downstream few-shot classification methods, even showing stronger performance on a set of challenging visual classification tasks.
Authors: Pascal J. Sager, Jan M. Deriu, Benjamin F. Grewe, Thilo Stadelmann, Christoph von der Malsburg
Abstract: We present a novel intelligent-system architecture called "Dynamic Net Architecture" (DNA) that relies on recurrence-stabilized networks and discuss it in application to vision. Our architecture models a (cerebral cortical) area wherein elementary feature neurons encode details of visual structures, and coherent nets of such neurons model holistic object structures. By interpreting smaller or larger coherent pieces of an area network as complex features, our model encodes hierarchical feature representations essentially different than artificial neural networks (ANNs). DNA models operate on a dynamic connectionism principle, wherein neural activations stemming from initial afferent signals undergo stabilization through a self-organizing mechanism facilitated by Hebbian plasticity alongside periodically tightening inhibition. In contrast to ANNs, which rely on feed-forward connections and backpropagation of error, we posit that this processing paradigm leads to highly robust representations, as by employing dynamic lateral connections, irrelevant details in neural activations are filtered out, freeing further processing steps from distracting noise and premature decisions. We empirically demonstrate the viability of the DNA by composing line fragments into longer lines and show that the construction of nets representing lines remains robust even with the introduction of up to $59\%$ noise at each spatial location. Furthermore, we demonstrate the model's capability to reconstruct anticipated features from partially obscured inputs and that it can generalize to patterns not observed during training. In this work, we limit the DNA to one cortical area and focus on its internals while providing insights into a standalone area's strengths and shortcomings. Additionally, we provide an outlook on how future work can implement invariant object recognition by combining multiple areas.
Authors: Fei Guo, YiKang Wang, Han Qi, Li Zhu, Jing Sun
Abstract: Few-shot action recognition is an emerging field in computer vision, primarily focused on meta-learning within the same domain. However, challenges arise in real-world scenario deployment, as gathering extensive labeled data within a specific domain is laborious and time-intensive. Thus, attention shifts towards cross-domain few-shot action recognition, requiring the model to generalize across domains with significant deviations. Therefore, we propose a novel approach, ``Distillation from Mixed-Source Domain", tailored to address this conundrum. Our method strategically integrates insights from both labeled data of the source domain and unlabeled data of the target domain during the training. The ResNet18 is used as the backbone to extract spatial features from the source and target domains. We design two branches for meta-training: the original-source and the mixed-source branches. In the first branch, a Domain Temporal Encoder is employed to capture temporal features for both the source and target domains. Additionally, a Domain Temporal Decoder is employed to reconstruct all extracted features. In the other branch, a Domain Mixed Encoder is used to handle labeled source domain data and unlabeled target domain data, generating mixed-source domain features. We incorporate a pre-training stage before meta-training, featuring a network architecture similar to that of the first branch. Lastly, we introduce a dual distillation mechanism to refine the classification probabilities of source domain features, aligning them with those of mixed-source domain features. This iterative process enriches the insights of the original-source branch with knowledge from the mixed-source branch, thereby enhancing the model's generalization capabilities. Our code is available at URL: \url{https://xxxx/xxxx/xxxx.git}
Authors: Jiawei Guo, HungChyun Chou, Ning Ding
Abstract: Neural Radiance Fields (NeRF) are an advanced technology that creates highly realistic images by learning about scenes through a neural network model. However, NeRF often encounters issues when there are not enough images to work with, leading to problems in accurately rendering views. The main issue is that NeRF lacks sufficient structural details to guide the rendering process accurately. To address this, we proposed a Depth and Normal Dense Completion Priors for NeRF (CP\_NeRF) framework. This framework enhances view rendering by adding depth and normal dense completion priors to the NeRF optimization process. Before optimizing NeRF, we obtain sparse depth maps using the Structure from Motion (SfM) technique used to get camera poses. Based on the sparse depth maps and a normal estimator, we generate sparse normal maps for training a normal completion prior with precise standard deviations. During optimization, we apply depth and normal completion priors to transform sparse data into dense depth and normal maps with their standard deviations. We use these dense maps to guide ray sampling, assist distance sampling and construct a normal loss function for better training accuracy. To improve the rendering of NeRF's normal outputs, we incorporate an optical centre position embedder that helps synthesize more accurate normals through volume rendering. Additionally, we employ a normal patch matching technique to choose accurate rendered normal maps, ensuring more precise supervision for the model. Our method is superior to leading techniques in rendering detailed indoor scenes, even with limited input views.
Authors: Zhanwen Liu, Chao Li, Nan Yang, Yang Wang, Jiaqi Ma, Guangliang Cheng, Xiangmo Zhao
Abstract: Motion forecasting plays a pivotal role in autonomous driving systems, enabling vehicles to execute collision warnings and rational local-path planning based on predictions of the surrounding vehicles. However, prevalent methods often assume complete observed trajectories, neglecting the potential impact of missing values induced by object occlusion, scope limitation, and sensor failures. Such oversights inevitably compromise the accuracy of trajectory predictions. To tackle this challenge, we propose an end-to-end framework, termed Multiscale Transformer (MSTF), meticulously crafted for incomplete trajectory prediction. MSTF integrates a Multiscale Attention Head (MAH) and an Information Increment-based Pattern Adaptive (IIPA) module. Specifically, the MAH component concurrently captures multiscale motion representation of trajectory sequence from various temporal granularities, utilizing a multi-head attention mechanism. This approach facilitates the modeling of global dependencies in motion across different scales, thereby mitigating the adverse effects of missing values. Additionally, the IIPA module adaptively extracts continuity representation of motion across time steps by analyzing missing patterns in the data. The continuity representation delineates motion trend at a higher level, guiding MSTF to generate predictions consistent with motion continuity. We evaluate our proposed MSTF model using two large-scale real-world datasets. Experimental results demonstrate that MSTF surpasses state-of-the-art (SOTA) models in the task of incomplete trajectory prediction, showcasing its efficacy in addressing the challenges posed by missing values in motion forecasting for autonomous driving systems.
Authors: Yumeng Zhang, Shi Gong, Kaixin Xiong, Xiaoqing Ye, Xiao Tan, Fan Wang, Jizhou Huang, Hua Wu, Haifeng Wang
Abstract: World models are receiving increasing attention in autonomous driving for their ability to predict potential future scenarios. In this paper, we present BEVWorld, a novel approach that tokenizes multimodal sensor inputs into a unified and compact Bird's Eye View (BEV) latent space for environment modeling. The world model consists of two parts: the multi-modal tokenizer and the latent BEV sequence diffusion model. The multi-modal tokenizer first encodes multi-modality information and the decoder is able to reconstruct the latent BEV tokens into LiDAR and image observations by ray-casting rendering in a self-supervised manner. Then the latent BEV sequence diffusion model predicts future scenarios given action tokens as conditions. Experiments demonstrate the effectiveness of BEVWorld in autonomous driving tasks, showcasing its capability in generating future scenes and benefiting downstream tasks such as perception and motion prediction. Code will be available at https://github.com/zympsyche/BevWorld.
Authors: Qijun Gan, Wentong Li, Jinwei Ren, Jianke Zhu
Abstract: Reconstructing high-fidelity hand models with intricate textures plays a crucial role in enhancing human-object interaction and advancing real-world applications. Despite the state-of-the-art methods excelling in texture generation and image rendering, they often face challenges in accurately capturing geometric details. Learning-based approaches usually offer better robustness and faster inference, which tend to produce smoother results and require substantial amounts of training data. To address these issues, we present a novel fine-grained multi-view hand mesh reconstruction method that leverages inverse rendering to restore hand poses and intricate details. Firstly, our approach predicts a parametric hand mesh model through Graph Convolutional Networks (GCN) based method from multi-view images. We further introduce a novel Hand Albedo and Mesh (HAM) optimization module to refine both the hand mesh and textures, which is capable of preserving the mesh topology. In addition, we suggest an effective mesh-based neural rendering scheme to simultaneously generate photo-realistic image and optimize mesh geometry by fusing the pre-trained rendering network with vertex features. We conduct the comprehensive experiments on InterHand2.6M, DeepHandMesh and dataset collected by ourself, whose promising results show that our proposed approach outperforms the state-of-the-art methods on both reconstruction accuracy and rendering quality. Code and dataset are publicly available at https://github.com/agnJason/FMHR.
Authors: Yuxiang Yang, Lu Wen, Xinyi Zeng, Yuanyuan Xu, Xi Wu, Jiliu Zhou, Yan Wang
Abstract: Facial Expression Recognition (FER) holds significant importance in human-computer interactions. Existing cross-domain FER methods often transfer knowledge solely from a single labeled source domain to an unlabeled target domain, neglecting the comprehensive information across multiple sources. Nevertheless, cross-multidomain FER (CMFER) is very challenging for (i) the inherent inter-domain shifts across multiple domains and (ii) the intra-domain shifts stemming from the ambiguous expressions and low inter-class distinctions. In this paper, we propose a novel Learning with Alignments CMFER framework, named LA-CMFER, to handle both inter- and intra-domain shifts. Specifically, LA-CMFER is constructed with a global branch and a local branch to extract features from the full images and local subtle expressions, respectively. Based on this, LA-CMFER presents a dual-level inter-domain alignment method to force the model to prioritize hard-to-align samples in knowledge transfer at a sample level while gradually generating a well-clustered feature space with the guidance of class attributes at a cluster level, thus narrowing the inter-domain shifts. To address the intra-domain shifts, LA-CMFER introduces a multi-view intra-domain alignment method with a multi-view clustering consistency constraint where a prediction similarity matrix is built to pursue consistency between the global and local views, thus refining pseudo labels and eliminating latent noise. Extensive experiments on six benchmark datasets have validated the superiority of our LA-CMFER.
Authors: Huihui Xu, Yijun Yang, Angelica I Aviles-Rivero, Guang Yang, Jing Qin, Lei Zhu
Abstract: Regular screening and early discovery of uterine fibroid are crucial for preventing potential malignant transformations and ensuring timely, life-saving interventions. To this end, we collect and annotate the first ultrasound video dataset with 100 videos for uterine fibroid segmentation (UFUV). We also present Local-Global Reciprocal Network (LGRNet) to efficiently and effectively propagate the long-term temporal context which is crucial to help distinguish between uninformative noisy surrounding tissues and target lesion regions. Specifically, the Cyclic Neighborhood Propagation (CNP) is introduced to propagate the inter-frame local temporal context in a cyclic manner. Moreover, to aggregate global temporal context, we first condense each frame into a set of frame bottleneck queries and devise Hilbert Selective Scan (HilbertSS) to both efficiently path connect each frame and preserve the locality bias. A distribute layer is then utilized to disseminate back the global context for reciprocal refinement. Extensive experiments on UFUV and three public Video Polyp Segmentation (VPS) datasets demonstrate consistent improvements compared to state-of-the-art segmentation methods, indicating the effectiveness and versatility of LGRNet. Code, checkpoints, and dataset are available at https://github.com/bio-mlhui/LGRNet
Authors: Jianwen Jiang, Gaojie Lin, Zhengkun Rong, Chao Liang, Yongming Zhu, Jiaqi Yang, Tianyun Zhong
Abstract: Existing neural head avatars methods have achieved significant progress in the image quality and motion range of portrait animation. However, these methods neglect the computational overhead, and to the best of our knowledge, none is designed to run on mobile devices. This paper presents MobilePortrait, a lightweight one-shot neural head avatars method that reduces learning complexity by integrating external knowledge into both the motion modeling and image synthesis, enabling real-time inference on mobile devices. Specifically, we introduce a mixed representation of explicit and implicit keypoints for precise motion modeling and precomputed visual features for enhanced foreground and background synthesis. With these two key designs and using simple U-Nets as backbones, our method achieves state-of-the-art performance with less than one-tenth the computational demand. It has been validated to reach speeds of over 100 FPS on mobile devices and support both video and audio-driven inputs.
Authors: Hyunjin Cho, Dong Un Kang, Se Young Chun
Abstract: Short-term object interaction anticipation is an important task in egocentric video analysis, including precise predictions of future interactions and their timings as well as the categories and positions of the involved active objects. To alleviate the complexity of this task, our proposed method, SOIA-DOD, effectively decompose it into 1) detecting active object and 2) classifying interaction and predicting their timing. Our method first detects all potential active objects in the last frame of egocentric video by fine-tuning a pre-trained YOLOv9. Then, we combine these potential active objects as query with transformer encoder, thereby identifying the most promising next active object and predicting its future interaction and time-to-contact. Experimental results demonstrate that our method outperforms state-of-the-art models on the challenge test set, achieving the best performance in predicting next active objects and their interactions. Finally, our proposed ranked the third overall top-5 mAP when including time-to-contact predictions. The source code is available at https://github.com/KeenyJin/SOIA-DOD.
Authors: Zirui Zhou, Junhao Liang, Zizhao Peng, Chao Fan, Fengwei An, Shiqi Yu
Abstract: Scoliosis poses significant diagnostic challenges, particularly in adolescents, where early detection is crucial for effective treatment. Traditional diagnostic and follow-up methods, which rely on physical examinations and radiography, face limitations due to the need for clinical expertise and the risk of radiation exposure, thus restricting their use for widespread early screening. In response, we introduce a novel, video-based, non-invasive method for scoliosis classification using gait analysis, which circumvents these limitations. This study presents Scoliosis1K, the first large-scale dataset tailored for video-based scoliosis classification, encompassing over one thousand adolescents. Leveraging this dataset, we developed ScoNet, an initial model that encountered challenges in dealing with the complexities of real-world data. This led to the creation of ScoNet-MT, an enhanced model incorporating multi-task learning, which exhibits promising diagnostic accuracy for application purposes. Our findings demonstrate that gait can be a non-invasive biomarker for scoliosis, revolutionizing screening practices with deep learning and setting a precedent for non-invasive diagnostic methodologies. The dataset and code are publicly available at https://zhouzi180.github.io/Scoliosis1K/.
Authors: Yaoyao Zhu, Xiuding Cai, Dong Miao, Yu Yao, Zhongliang Fu
Abstract: Deep models often struggle with out-of-distribution (OOD) generalization, limiting their real-world applicability beyond controlled laboratory settings. Invariant risk minimization (IRM) addresses this issue by learning invariant features and minimizing the risk across different domains. Thus, it avoids the pitfalls of pseudo-invariant features and spurious causality associated with empirical risk minimization (ERM). However, according to the support overlap theorem, ERM and IRM may fail to address the OOD problem when pseudo-invariant features have insufficient support overlap. To this end, we propose a novel method to enlarge feature support overlap for domain generalization. Specifically, we introduce Bayesian random semantic data augmentation to increase sample diversity and overcome the deficiency of IRM. Experiments on several challenging OOD generalization benchmarks demonstrate that our approach surpasses existing models, delivering superior performance and robustness. The code is available at \url{https://github.com/YaoyaoZhu19/BSDG}.
Authors: Hao Jing, Anhong Wang, Lijun Zhao, Yakun Yang, Donghan Bu, Jing Zhang, Yifan Zhang, Junhui Hou
Abstract: In autonomous driving, LiDAR sensors are vital for acquiring 3D point clouds, providing reliable geometric information. However, traditional sampling methods of preprocessing often ignore semantic features, leading to detail loss and ground point interference in 3D object detection. To address this, we propose a multi-branch two-stage 3D object detection framework using a Semantic-aware Multi-branch Sampling (SMS) module and multi-view consistency constraints. The SMS module includes random sampling, Density Equalization Sampling (DES) for enhancing distant objects, and Ground Abandonment Sampling (GAS) to focus on non-ground points. The sampled multi-view points are processed through a Consistent KeyPoint Selection (CKPS) module to generate consistent keypoint masks for efficient proposal sampling. The first-stage detector uses multi-branch parallel learning with multi-view consistency loss for feature aggregation, while the second-stage detector fuses multi-view data through a Multi-View Fusion Pooling (MVFP) module to precisely predict 3D objects. The experimental results on KITTI 3D object detection benchmark dataset show that our method achieves excellent detection performance improvement for a variety of backbones, especially for low-performance backbones with the simple network structures.
Authors: Tengjie Zhu, Zhuo Chen, Jingnan Gao, Yichao Yan, Xiaokang Yang
Abstract: Inverse rendering methods have achieved remarkable performance in reconstructing high-fidelity 3D objects with disentangled geometries, materials, and environmental light. However, they still face huge challenges in reflective surface reconstruction. Although recent methods model the light trace to learn specularity, the ignorance of indirect illumination makes it hard to handle inter-reflections among multiple smooth objects. In this work, we propose Ref-MC2 that introduces the multi-time Monte Carlo sampling which comprehensively computes the environmental illumination and meanwhile considers the reflective light from object surfaces. To address the computation challenge as the times of Monte Carlo sampling grow, we propose a specularity-adaptive sampling strategy, significantly reducing the computational complexity. Besides the computational resource, higher geometry accuracy is also required because geometric errors accumulate multiple times. Therefore, we further introduce a reflection-aware surface model to initialize the geometry and refine it during inverse rendering. We construct a challenging dataset containing scenes with multiple objects and inter-reflections. Experiments show that our method outperforms other inverse rendering methods on various object groups. We also show downstream applications, e.g., relighting and material editing, to illustrate the disentanglement ability of our method.
Authors: Yingying Jiang, Hanchao Jia, Xiaobing Wang, Peng Hao
Abstract: Composed Image Retrieval (CIR) aims to retrieve images based on a query image with text. Current Zero-Shot CIR (ZS-CIR) methods try to solve CIR tasks without using expensive triplet-labeled training datasets. However, the gap between ZS-CIR and triplet-supervised CIR is still large. In this work, we propose Hybrid CIR (HyCIR), which uses synthetic labels to boost the performance of ZS-CIR. A new label Synthesis pipeline for CIR (SynCir) is proposed, in which only unlabeled images are required. First, image pairs are extracted based on visual similarity. Second, query text is generated for each image pair based on vision-language model and LLM. Third, the data is further filtered in language space based on semantic similarity. To improve ZS-CIR performance, we propose a hybrid training strategy to work with both ZS-CIR supervision and synthetic CIR triplets. Two kinds of contrastive learning are adopted. One is to use large-scale unlabeled image dataset to learn an image-to-text mapping with good generalization. The other is to use synthetic CIR triplets to learn a better mapping for CIR tasks. Our approach achieves SOTA zero-shot performance on the common CIR benchmarks: CIRR and CIRCO.
Authors: Sushil Sharma, Arindam Das, Ganesh Sistu, Mark Halton, Ciar\'an Eising
Abstract: Predicting ego vehicle trajectories remains a critical challenge, especially in urban and dense areas due to the unpredictable behaviours of other vehicles and pedestrians. Multimodal trajectory prediction enhances decision-making by considering multiple possible future trajectories based on diverse sources of environmental data. In this approach, we leverage ResNet-50 to extract image features from high-definition map data and use IMU sensor data to calculate speed, acceleration, and yaw rate. A temporal probabilistic network is employed to compute potential trajectories, selecting the most accurate and highly probable trajectory paths. This method integrates HD map data to improve the robustness and reliability of trajectory predictions for autonomous vehicles.
Authors: Yaozong Gan, Guang Li, Ren Togo, Keisuke Maeda, Takahiro Ogawa, Miki Haseyama
Abstract: Recent multimodal large language models (MLLM) such as GPT-4o and GPT-4v have shown great potential in autonomous driving. In this paper, we propose a cross-domain few-shot in-context learning method based on the MLLM for enhancing traffic sign recognition (TSR). We first construct a traffic sign detection network based on Vision Transformer Adapter and an extraction module to extract traffic signs from the original road images. To reduce the dependence on training data and improve the performance stability of cross-country TSR, we introduce a cross-domain few-shot in-context learning method based on the MLLM. To enhance MLLM's fine-grained recognition ability of traffic signs, the proposed method generates corresponding description texts using template traffic signs. These description texts contain key information about the shape, color, and composition of traffic signs, which can stimulate the ability of MLLM to perceive fine-grained traffic sign categories. By using the description texts, our method reduces the cross-domain differences between template and real traffic signs. Our approach requires only simple and uniform textual indications, without the need for large-scale traffic sign images and labels. We perform comprehensive evaluations on the German traffic sign recognition benchmark dataset, the Belgium traffic sign dataset, and two real-world datasets taken from Japan. The experimental results show that our method significantly enhances the TSR performance.
Authors: Chinmay Prabhakar, Suprosanna Shit, Fabio Musio, Kaiyuan Yang, Tamaz Amiranashvili, Johannes C. Paetzold, Hongwei Bran Li, Bjoern Menze
Abstract: Blood vessel networks, represented as 3D graphs, help predict disease biomarkers, simulate blood flow, and aid in synthetic image generation, relevant in both clinical and pre-clinical settings. However, generating realistic vessel graphs that correspond to an anatomy of interest is challenging. Previous methods aimed at generating vessel trees mostly in an autoregressive style and could not be applied to vessel graphs with cycles such as capillaries or specific anatomical structures such as the Circle of Willis. Addressing this gap, we introduce the first application of \textit{denoising diffusion models} in 3D vessel graph generation. Our contributions include a novel, two-stage generation method that sequentially denoises node coordinates and edges. We experiment with two real-world vessel datasets, consisting of microscopic capillaries and major cerebral vessels, and demonstrate the generalizability of our method for producing diverse, novel, and anatomically plausible vessel graphs.
Authors: Kaouther Mouheb, Marawan Elbatel, Stefan Klein, Esther E. Bron
Abstract: Deep learning has achieved impressive performance across various medical imaging tasks. However, its inherent bias against specific groups hinders its clinical applicability in equitable healthcare systems. A recently discovered phenomenon, Neural Collapse (NC), has shown potential in improving the generalization of state-of-the-art deep learning models. Nonetheless, its implications on bias in medical imaging remain unexplored. Our study investigates deep learning fairness through the lens of NC. We analyze the training dynamics of models as they approach NC when training using biased datasets, and examine the subsequent impact on test performance, specifically focusing on label bias. We find that biased training initially results in different NC configurations across subgroups, before converging to a final NC solution by memorizing all data samples. Through extensive experiments on three medical imaging datasets -- PAPILA, HAM10000, and CheXpert -- we find that in biased settings, NC can lead to a significant drop in F1 score across all subgroups. Our code is available at https://gitlab.com/radiology/neuro/neural-collapse-fairness
URLs: https://gitlab.com/radiology/neuro/neural-collapse-fairness
Authors: Alexander Jaus, Constantin Seibold, Simon Rei{\ss}, Lukas Heine, Anton Schily, Moon Kim, Fin Hendrik Bahnsen, Ken Herrmann, Rainer Stiefelhagen, Jens Kleesiek
Abstract: Pathological structures in medical images are typically deviations from the expected anatomy of a patient. While clinicians consider this interplay between anatomy and pathology, recent deep learning algorithms specialize in recognizing either one of the two, rarely considering the patient's body from such a joint perspective. In this paper, we develop a generalist segmentation model that combines anatomical and pathological information, aiming to enhance the segmentation accuracy of pathological features. Our Anatomy-Pathology Exchange (APEx) training utilizes a query-based segmentation transformer which decodes a joint feature space into query-representations for human anatomy and interleaves them via a mixing strategy into the pathology-decoder for anatomy-informed pathology predictions. In doing so, we are able to report the best results across the board on FDG-PET-CT and Chest X-Ray pathology segmentation tasks with a margin of up to 3.3% as compared to strong baseline methods. Code and models will be publicly available at github.com/alexanderjaus/APEx.
Authors: Shahaf E. Finder, Roy Amoyal, Eran Treister, Oren Freifeld
Abstract: In recent years, there have been attempts to increase the kernel size of Convolutional Neural Nets (CNNs) to mimic the global receptive field of Vision Transformers' (ViTs) self-attention blocks. That approach, however, quickly hit an upper bound and saturated way before achieving a global receptive field. In this work, we demonstrate that by leveraging the Wavelet Transform (WT), it is, in fact, possible to obtain very large receptive fields without suffering from over-parameterization, e.g., for a $k \times k$ receptive field, the number of trainable parameters in the proposed method grows only logarithmically with $k$. The proposed layer, named WTConv, can be used as a drop-in replacement in existing architectures, results in an effective multi-frequency response, and scales gracefully with the size of the receptive field. We demonstrate the effectiveness of the WTConv layer within ConvNeXt and MobileNetV2 architectures for image classification, as well as backbones for downstream tasks, and show it yields additional properties such as robustness to image corruption and an increased response to shapes over textures. Our code is available at https://github.com/BGU-CS-VIL/WTConv.
Authors: Bin Ren, Guofeng Mei, Danda Pani Paudel, Weijie Wang, Yawei Li, Mengyuan Liu, Rita Cucchiara, Luc Van Gool, Nicu Sebe
Abstract: Contrastive learning (CL) for Vision Transformers (ViTs) in image domains has achieved performance comparable to CL for traditional convolutional backbones. However, in 3D point cloud pretraining with ViTs, masked autoencoder (MAE) modeling remains dominant. This raises the question: Can we take the best of both worlds? To answer this question, we first empirically validate that integrating MAE-based point cloud pre-training with the standard contrastive learning paradigm, even with meticulous design, can lead to a decrease in performance. To address this limitation, we reintroduce CL into the MAE-based point cloud pre-training paradigm by leveraging the inherent contrastive properties of MAE. Specifically, rather than relying on extensive data augmentation as commonly used in the image domain, we randomly mask the input tokens twice to generate contrastive input pairs. Subsequently, a weight-sharing encoder and two identically structured decoders are utilized to perform masked token reconstruction. Additionally, we propose that for an input token masked by both masks simultaneously, the reconstructed features should be as similar as possible. This naturally establishes an explicit contrastive constraint within the generative MAE-based pre-training paradigm, resulting in our proposed method, Point-CMAE. Consequently, Point-CMAE effectively enhances the representation quality and transfer performance compared to its MAE counterpart. Experimental evaluations across various downstream applications, including classification, part segmentation, and few-shot learning, demonstrate the efficacy of our framework in surpassing state-of-the-art techniques under standard ViTs and single-modal settings. The source code and trained models are available at: https://github.com/Amazingren/Point-CMAE.
Authors: Lintao Zhang, Xiangcheng Du, LeoWu TomyEnrique, Yiqun Wang, Yingbin Zheng, Cheng Jin
Abstract: For image inpainting, the existing Denoising Diffusion Probabilistic Model (DDPM) based method i.e. RePaint can produce high-quality images for any inpainting form. It utilizes a pre-trained DDPM as a prior and generates inpainting results by conditioning on the reverse diffusion process, namely denoising process. However, this process is significantly time-consuming. In this paper, we propose an efficient DDPM-based image inpainting method which includes three speed-up strategies. First, we utilize a pre-trained Light-Weight Diffusion Model (LWDM) to reduce the number of parameters. Second, we introduce a skip-step sampling scheme of Denoising Diffusion Implicit Models (DDIM) for the denoising process. Finally, we propose Coarse-to-Fine Sampling (CFS), which speeds up inference by reducing image resolution in the coarse stage and decreasing denoising timesteps in the refinement stage. We conduct extensive experiments on both faces and general-purpose image inpainting tasks, and our method achieves competitive performance with approximately 60 times speedup.
Authors: Xiang Zhang, Yulun Zhang, Fisher Yu
Abstract: Transformers have exhibited promising performance in computer vision tasks including image super-resolution (SR). However, popular transformer-based SR methods often employ window self-attention with quadratic computational complexity to window sizes, resulting in fixed small windows with limited receptive fields. In this paper, we present a general strategy to convert transformer-based SR networks to hierarchical transformers (HiT-SR), boosting SR performance with multi-scale features while maintaining an efficient design. Specifically, we first replace the commonly used fixed small windows with expanding hierarchical windows to aggregate features at different scales and establish long-range dependencies. Considering the intensive computation required for large windows, we further design a spatial-channel correlation method with linear complexity to window sizes, efficiently gathering spatial and channel information from hierarchical windows. Extensive experiments verify the effectiveness and efficiency of our HiT-SR, and our improved versions of SwinIR-Light, SwinIR-NG, and SRFormer-Light yield state-of-the-art SR results with fewer parameters, FLOPs, and faster speeds ($\sim7\times$).
Authors: Reza Abbasi, Mohammad Hossein Rohban, Mahdieh Soleymani Baghshah
Abstract: CLIP models have recently shown to exhibit Out of Distribution (OoD) generalization capabilities. However, Compositional Out of Distribution (C-OoD) generalization, which is a crucial aspect of a model's ability to understand unseen compositions of known concepts, is relatively unexplored for the CLIP models. Our goal is to address this problem and identify the factors that contribute to the C-OoD in CLIPs. We noted that previous studies regarding compositional understanding of CLIPs frequently fail to ensure that test samples are genuinely novel relative to the CLIP training data. To this end, we carefully synthesized a large and diverse dataset in the single object setting, comprising attributes for objects that are highly unlikely to be encountered in the combined training datasets of various CLIP models. This dataset enables an authentic evaluation of C-OoD generalization. Our observations reveal varying levels of C-OoD generalization across different CLIP models. We propose that the disentanglement of CLIP representations serves as a critical indicator in this context. By utilizing our synthesized datasets and other existing datasets, we assess various disentanglement metrics of text and image representations. Our study reveals that the disentanglement of image and text representations, particularly with respect to their compositional elements, plays a crucial role in improving the generalization of CLIP models in out-of-distribution settings. This finding suggests promising opportunities for advancing out-of-distribution generalization in CLIPs.
Authors: Chenxu Wang, Chunyan Xu, Ziqi Gu, Zhen Cui
Abstract: While existing semi-supervised object detection (SSOD) methods perform well in general scenes, they encounter challenges in handling oriented objects in aerial images. We experimentally find three gaps between general and oriented object detection in semi-supervised learning: 1) Sampling inconsistency: the common center sampling is not suitable for oriented objects with larger aspect ratios when selecting positive labels from labeled data. 2) Assignment inconsistency: balancing the precision and localization quality of oriented pseudo-boxes poses greater challenges which introduces more noise when selecting positive labels from unlabeled data. 3) Confidence inconsistency: there exists more mismatch between the predicted classification and localization qualities when considering oriented objects, affecting the selection of pseudo-labels. Therefore, we propose a Multi-clue Consistency Learning (MCL) framework to bridge gaps between general and oriented objects in semi-supervised detection. Specifically, considering various shapes of rotated objects, the Gaussian Center Assignment is specially designed to select the pixel-level positive labels from labeled data. We then introduce the Scale-aware Label Assignment to select pixel-level pseudo-labels instead of unreliable pseudo-boxes, which is a divide-and-rule strategy suited for objects with various scales. The Consistent Confidence Soft Label is adopted to further boost the detector by maintaining the alignment of the predicted results. Comprehensive experiments on DOTA-v1.5 and DOTA-v1.0 benchmarks demonstrate that our proposed MCL can achieve state-of-the-art performance in the semi-supervised oriented object detection task.
Authors: Aaron Lohner, Francesco Compagno, Jonathan Francis, Alessandro Oltramari
Abstract: Recognizing a traffic accident is an essential part of any autonomous driving or road monitoring system. An accident can appear in a wide variety of forms, and understanding what type of accident is taking place may be useful to prevent it from reoccurring. The task of being able to classify a traffic scene as a specific type of accident is the focus of this work. We approach the problem by likening a traffic scene to a graph, where objects such as cars can be represented as nodes, and relative distances and directions between them as edges. This representation of an accident can be referred to as a scene graph, and is used as input for an accident classifier. Better results can be obtained with a classifier that fuses the scene graph input with representations from vision and language. This work introduces a multi-stage, multimodal pipeline to pre-process videos of traffic accidents, encode them as scene graphs, and align this representation with vision and language modalities for accident classification. When trained on 4 classes, our method achieves a balanced accuracy score of 57.77% on an (unbalanced) subset of the popular Detection of Traffic Anomaly (DoTA) benchmark, representing an increase of close to 5 percentage points from the case where scene graph information is not taken into account.
Authors: Tinghuai Wang
Abstract: Learning a data-driven spatio-temporal semantic representation of the objects is the key to coherent and consistent labelling in video. This paper proposes to achieve semantic video object segmentation by learning a data-driven representation which captures the synergy of multiple instances from continuous frames. To prune the noisy detections, we exploit the rich information among multiple instances and select the discriminative and representative subset. This selection process is formulated as a facility location problem solved by maximising a submodular function. Our method retrieves the longer term contextual dependencies which underpins a robust semantic video object segmentation algorithm. We present extensive experiments on a challenging dataset that demonstrate the superior performance of our approach compared with the state-of-the-art methods.
Authors: Tinghuai Wang, Huiling Wang
Abstract: We propose a novel approach for modeling semantic contextual relationships in videos. This graph-based model enables the learning and propagation of higher-level spatial-temporal contexts to facilitate the semantic labeling of local regions. We introduce an exemplar-based nonparametric view of contextual cues, where the inherent relationships implied by object hypotheses are encoded on a similarity graph of regions. Contextual relationships learning and propagation are performed to estimate the pairwise contexts between all pairs of unlabeled local regions. Our algorithm integrates the learned contexts into a Conditional Random Field (CRF) in the form of pairwise potentials and infers the per-region semantic labels. We evaluate our approach on the challenging YouTube-Objects dataset which shows that the proposed contextual relationship model outperforms the state-of-the-art methods.
Authors: Skanda Koppula, Ignacio Rocco, Yi Yang, Joe Heyward, Jo\~ao Carreira, Andrew Zisserman, Gabriel Brostow, Carl Doersch
Abstract: We introduce a new benchmark, TAPVid-3D, for evaluating the task of long-range Tracking Any Point in 3D (TAP-3D). While point tracking in two dimensions (TAP) has many benchmarks measuring performance on real-world videos, such as TAPVid-DAVIS, three-dimensional point tracking has none. To this end, leveraging existing footage, we build a new benchmark for 3D point tracking featuring 4,000+ real-world videos, composed of three different data sources spanning a variety of object types, motion patterns, and indoor and outdoor environments. To measure performance on the TAP-3D task, we formulate a collection of metrics that extend the Jaccard-based metric used in TAP to handle the complexities of ambiguous depth scales across models, occlusions, and multi-track spatio-temporal smoothness. We manually verify a large sample of trajectories to ensure correct video annotations, and assess the current state of the TAP-3D task by constructing competitive baselines using existing tracking models. We anticipate this benchmark will serve as a guidepost to improve our ability to understand precise 3D motion and surface deformation from monocular video. Code for dataset download, generation, and model evaluation is available at https://tapvid3d.github.io
Authors: Tinghuai Wang, Huiling Wang
Abstract: Human body parsing remains a challenging problem in natural scenes due to multi-instance and inter-part semantic confusions as well as occlusions. This paper proposes a novel approach to decomposing multiple human bodies into semantic part regions in unconstrained environments. Specifically we propose a convolutional neural network (CNN) architecture which comprises of novel semantic and contour attention mechanisms across feature hierarchy to resolve the semantic ambiguities and boundary localization issues related to semantic body parsing. We further propose to encode estimated pose as higher-level contextual information which is combined with local semantic cues in a novel graphical model in a principled manner. In this proposed model, the lower-level semantic cues can be recursively updated by propagating higher-level contextual information from estimated pose and vice versa across the graph, so as to alleviate erroneous pose information and pixel level predictions. We further propose an optimization technique to efficiently derive the solutions. Our proposed method achieves the state-of-art results on the challenging Pascal Person-Part dataset.
Authors: Yibo Miao, Yifan Zhu, Yinpeng Dong, Lijia Yu, Jun Zhu, Xiao-Shan Gao
Abstract: The recent development of Sora leads to a new era in text-to-video (T2V) generation. Along with this comes the rising concern about its security risks. The generated videos may contain illegal or unethical content, and there is a lack of comprehensive quantitative understanding of their safety, posing a challenge to their reliability and practical deployment. Previous evaluations primarily focus on the quality of video generation. While some evaluations of text-to-image models have considered safety, they cover fewer aspects and do not address the unique temporal risk inherent in video generation. To bridge this research gap, we introduce T2VSafetyBench, a new benchmark designed for conducting safety-critical assessments of text-to-video models. We define 12 critical aspects of video generation safety and construct a malicious prompt dataset using LLMs and jailbreaking prompt attacks. Based on our evaluation results, we draw several important findings, including: 1) no single model excels in all aspects, with different models showing various strengths; 2) the correlation between GPT-4 assessments and manual reviews is generally high; 3) there is a trade-off between the usability and safety of text-to-video generative models. This indicates that as the field of video generation rapidly advances, safety risks are set to surge, highlighting the urgency of prioritizing video safety. We hope that T2VSafetyBench can provide insights for better understanding the safety of video generation in the era of generative AI.
Authors: Huilong Xie, Wenwei Song, Wenxiong Kang, Yihong Lin
Abstract: Recent advancements in both transformer-based methods and spiral neighbor sampling techniques have greatly enhanced hand mesh reconstruction. Transformers excel in capturing complex vertex relationships, and spiral neighbor sampling is vital for utilizing topological structures. This paper ingeniously integrates spiral sampling into the Transformer architecture, enhancing its ability to leverage mesh topology for superior performance in hand mesh reconstruction, resulting in substantial accuracy boosts. STMR employs a single image encoder for model efficiency. To augment its information extraction capability, we design the multi-scale pose feature extraction (MSPFE) module, which facilitates the extraction of rich pose features, ultimately enhancing the model's performance. Moreover, the proposed predefined pose-to-vertex lifting (PPVL) method improves vertex feature representation, further boosting reconstruction performance. Extensive experiments on the FreiHAND dataset demonstrate the state-of-the-art performance and unparalleled inference speed of STMR compared with similar backbone methods, showcasing its efficiency and effectiveness. The code is available at https://github.com/SmallXieGithub/STMR.
Authors: Zexin Ji, Beiji Zou, Xiaoyan Kui, Pierre Vera, Su Ruan
Abstract: In this paper, we propose a new architecture, called Deform-Mamba, for MR image super-resolution. Unlike conventional CNN or Transformer-based super-resolution approaches which encounter challenges related to the local respective field or heavy computational cost, our approach aims to effectively explore the local and global information of images. Specifically, we develop a Deform-Mamba encoder which is composed of two branches, modulated deform block and vision Mamba block. We also design a multi-view context module in the bottleneck layer to explore the multi-view contextual content. Thanks to the extracted features of the encoder, which include content-adaptive local and efficient global information, the vision Mamba decoder finally generates high-quality MR images. Moreover, we introduce a contrastive edge loss to promote the reconstruction of edge and contrast related content. Quantitative and qualitative experimental results indicate that our approach on IXI and fastMRI datasets achieves competitive performance.
Authors: Hozaifa Kassab, Ahmed Mahmoud, Mohamed Bahaa, Ammar Mohamed, Ali Hamdi
Abstract: We introduce MMIS, a novel dataset designed to advance MultiModal Interior Scene generation and recognition. MMIS consists of nearly 160,000 images. Each image within the dataset is accompanied by its corresponding textual description and an audio recording of that description, providing rich and diverse sources of information for scene generation and recognition. MMIS encompasses a wide range of interior spaces, capturing various styles, layouts, and furnishings. To construct this dataset, we employed careful processes involving the collection of images, the generation of textual descriptions, and corresponding speech annotations. The presented dataset contributes to research in multi-modal representation learning tasks such as image generation, retrieval, captioning, and classification.
Authors: Yuhang Lu, Zewei Xu, Touradj Ebrahimi
Abstract: Over recent years, deep convolutional neural networks have significantly advanced the field of face recognition techniques for both verification and identification purposes. Despite the impressive accuracy, these neural networks are often criticized for lacking explainability. There is a growing demand for understanding the decision-making process of AI-based face recognition systems. Some studies have investigated the use of visual saliency maps as explanations, but they have predominantly focused on the specific face verification case. The discussion on more general face recognition scenarios and the corresponding evaluation methodology for these explanations have long been absent in current research. Therefore, this manuscript conceives a comprehensive explanation framework for face recognition tasks. Firstly, an exhaustive definition of visual saliency map-based explanations for AI-based face recognition systems is provided, taking into account the two most common recognition situations individually, i.e., face verification and identification. Secondly, a new model-agnostic explanation method named CorrRISE is proposed to produce saliency maps, which reveal both the similar and dissimilar regions between any given face images. Subsequently, the explanation framework conceives a new evaluation methodology that offers quantitative measurement and comparison of the performance of general visual saliency explanation methods in face recognition. Consequently, extensive experiments are carried out on multiple verification and identification scenarios. The results showcase that CorrRISE generates insightful saliency maps and demonstrates superior performance, particularly in similarity maps in comparison with the state-of-the-art explanation approaches.
Authors: Makkunda Sharma, Fan Yang, Duy-Nhat Vo, Esra Suel, Swapnil Mishra, Samir Bhatt, Oliver Fiala, William Rudgard, Seth Flaxman
Abstract: Satellite imagery has emerged as an important tool to analyse demographic, health, and development indicators. While various deep learning models have been built for these tasks, each is specific to a particular problem, with few standard benchmarks available. We propose a new dataset pairing satellite imagery and high-quality survey data on child poverty to benchmark satellite feature representations. Our dataset consists of 33,608 images, each 10 km $\times$ 10 km, from 19 countries in Eastern and Southern Africa in the time period 1997-2022. As defined by UNICEF, multidimensional child poverty covers six dimensions and it can be calculated from the face-to-face Demographic and Health Surveys (DHS) Program . As part of the benchmark, we test spatial as well as temporal generalization, by testing on unseen locations, and on data after the training years. Using our dataset we benchmark multiple models, from low-level satellite imagery models such as MOSAIKS , to deep learning foundation models, which include both generic vision models such as Self-Distillation with no Labels (DINOv2) models and specific satellite imagery models such as SatMAE. We provide open source code for building the satellite dataset, obtaining ground truth data from DHS and running various models assessed in our work.
Authors: Zexin Ji, Beiji Zou, Xiaoyan Kui, Pierre Vera, Su Ruan
Abstract: In this paper, we propose a self-prior guided Mamba-UNet network (SMamba-UNet) for medical image super-resolution. Existing methods are primarily based on convolutional neural networks (CNNs) or Transformers. CNNs-based methods fail to capture long-range dependencies, while Transformer-based approaches face heavy calculation challenges due to their quadratic computational complexity. Recently, State Space Models (SSMs) especially Mamba have emerged, capable of modeling long-range dependencies with linear computational complexity. Inspired by Mamba, our approach aims to learn the self-prior multi-scale contextual features under Mamba-UNet networks, which may help to super-resolve low-resolution medical images in an efficient way. Specifically, we obtain self-priors by perturbing the brightness inpainting of the input image during network training, which can learn detailed texture and brightness information that is beneficial for super-resolution. Furthermore, we combine Mamba with Unet network to mine global features at different levels. We also design an improved 2D-Selective-Scan (ISS2D) module to divide image features into different directional sequences to learn long-range dependencies in multiple directions, and adaptively fuse sequence information to enhance super-resolved feature representation. Both qualitative and quantitative experimental results demonstrate that our approach outperforms current state-of-the-art methods on two public medical datasets: the IXI and fastMRI.
Authors: Mia Siemon, Thomas B. Moeslund, Barry Norton, Kamal Nasrollahi
Abstract: In this study, we formulate the task of Video Anomaly Detection as a probabilistic analysis of object bounding boxes. We hypothesize that the representation of objects via their bounding boxes only, can be sufficient to successfully identify anomalous events in a scene. The implied value of this approach is increased object anonymization, faster model training and fewer computational resources. This can particularly benefit applications within video surveillance running on edge devices such as cameras. We design our model based on human reasoning which lends itself to explaining model output in human-understandable terms. Meanwhile, the slowest model trains within less than 7 seconds on a 11th Generation Intel Core i9 Processor. While our approach constitutes a drastic reduction of problem feature space in comparison with prior art, we show that this does not result in a reduction in performance: the results we report are highly competitive on the benchmark datasets CUHK Avenue and ShanghaiTech, and significantly exceed on the latest State-of-the-Art results on StreetScene, which has so far proven to be the most challenging VAD dataset.
Authors: Bohan Hou, Haoqiang Lin, Haokun Wen, Meng Liu, Xuemeng Song
Abstract: Composed Image Retrieval (CIR) is a challenging task that aims to retrieve the target image based on a multimodal query, i.e., a reference image and its corresponding modification text. While previous supervised or zero-shot learning paradigms all fail to strike a good trade-off between time-consuming annotation cost and retrieval performance, recent researchers introduced the task of few-shot CIR (FS-CIR) and proposed a textual inversion-based network based on pretrained CLIP model to realize it. Despite its promising performance, the approach suffers from two key limitations: insufficient multimodal query composition training and indiscriminative training triplet selection. To address these two limitations, in this work, we propose a novel two-stage pseudo triplet guided few-shot CIR scheme, dubbed PTG-FSCIR. In the first stage, we employ a masked training strategy and advanced image caption generator to construct pseudo triplets from pure image data to enable the model to acquire primary knowledge related to multimodal query composition. In the second stage, based on active learning, we design a pseudo modification text-based query-target distance metric to evaluate the challenging score for each unlabeled sample. Meanwhile, we propose a robust top range-based random sampling strategy according to the 3-$\sigma$ rule in statistics, to sample the challenging samples for fine-tuning the pretrained model. Notably, our scheme is plug-and-play and compatible with any existing supervised CIR models. We tested our scheme across three backbones on three public datasets (i.e., FashionIQ, CIRR, and Birds-to-Words), achieving maximum improvements of 26.4%, 25.5% and 21.6% respectively, demonstrating our scheme's effectiveness.
Authors: Mohamed Bahaa, Mena Hany, Ehab E. Zakaria
Abstract: With the exponential increase in video content, the need for accurate deception detection in human-centric video analysis has become paramount. This research focuses on the extraction and combination of various features to enhance the accuracy of deception detection models. By systematically extracting features from visual, audio, and text data, and experimenting with different combinations, we developed a robust model that achieved an impressive 99% accuracy. Our methodology emphasizes the significance of feature engineering in deception detection, providing a clear and interpretable framework. We trained various machine learning models, including LSTM, BiLSTM, and pre-trained CNNs, using both single and multi-modal approaches. The results demonstrated that combining multiple modalities significantly enhances detection performance compared to single modality training. This study highlights the potential of strategic feature extraction and combination in developing reliable and transparent automated deception detection systems in video analysis, paving the way for more advanced and accurate detection methodologies in future research.
Authors: Sarah Elmahdy, Rodaina Hebishy, Ali Hamdi
Abstract: Night time semantic segmentation is a crucial task in computer vision, focusing on accurately classifying and segmenting objects in low-light conditions. Unlike daytime techniques, which often perform worse in nighttime scenes, it is essential for autonomous driving due to insufficient lighting, low illumination, dynamic lighting, shadow effects, and reduced contrast. We propose RHRSegNet, implementing a relighting model over a High-Resolution Network for semantic segmentation. RHRSegNet implements residual convolutional feature learning to handle complex lighting conditions. Our model then feeds the lightened scene feature maps into a high-resolution network for scene segmentation. The network consists of a convolutional producing feature maps with varying resolutions, achieving different levels of resolution through down-sampling and up-sampling. Large nighttime datasets are used for training and evaluation, such as NightCity, City-Scape, and Dark-Zurich datasets. Our proposed model increases the HRnet segmentation performance by 5% in low-light or nighttime images.
Authors: Shakeeb Murtaza, Marco Pedersoli, Aydin Sarraf, Eric Granger
Abstract: Weakly-Supervised Video Object Localization (WSVOL) involves localizing an object in videos using only video-level labels, also referred to as tags. State-of-the-art WSVOL methods like Temporal CAM (TCAM) rely on class activation mapping (CAM) and typically require a pre-trained CNN classifier. However, their localization accuracy is affected by their tendency to minimize the mutual information between different instances of a class and exploit temporal information during training for downstream tasks, e.g., detection and tracking. In the absence of bounding box annotation, it is challenging to exploit precise information about objects from temporal cues because the model struggles to locate objects over time. To address these issues, a novel method called transformer based CAM for videos (TrCAM-V), is proposed for WSVOL. It consists of a DeiT backbone with two heads for classification and localization. The classification head is trained using standard classification loss (CL), while the localization head is trained using pseudo-labels that are extracted using a pre-trained CLIP model. From these pseudo-labels, the high and low activation values are considered to be foreground and background regions, respectively. Our TrCAM-V method allows training a localization network by sampling pseudo-pixels on the fly from these regions. Additionally, a conditional random field (CRF) loss is employed to align the object boundaries with the foreground map. During inference, the model can process individual frames for real-time localization applications. Extensive experiments on challenging YouTube-Objects unconstrained video datasets show that our TrCAM-V method achieves new state-of-the-art performance in terms of classification and localization accuracy.
Authors: Puzuo Wang, Wei Yao, Jie Shao, Zhiyi He
Abstract: Domain adaptation (DA) techniques help deep learning models generalize across data shifts for point cloud semantic segmentation (PCSS). Test-time adaptation (TTA) allows direct adaptation of a pre-trained model to unlabeled data during inference stage without access to source data or additional training, avoiding privacy issues and large computational resources. We address TTA for geospatial PCSS by introducing three domain shift paradigms: photogrammetric to airborne LiDAR, airborne to mobile LiDAR, and synthetic to mobile laser scanning. We propose a TTA method that progressively updates batch normalization (BN) statistics with each testing batch. Additionally, a self-supervised learning module optimizes learnable BN affine parameters. Information maximization and reliability-constrained pseudo-labeling improve prediction confidence and supply supervisory signals. Experimental results show our method improves classification accuracy by up to 20\% mIoU, outperforming other methods. For photogrammetric (SensatUrban) to airborne (Hessigheim 3D) adaptation at the inference stage, our method achieves 59.46\% mIoU and 85.97\% OA without retraining or fine-turning.
Authors: Wenjun Miao, Guansong Pang, Trong-Tung Nguyen, Ruohang Fang, Jin Zheng, Xiao Bai
Abstract: Class incremental learning (CIL) aims to learn a model that can not only incrementally accommodate new classes, but also maintain the learned knowledge of old classes. Out-of-distribution (OOD) detection in CIL is to retain this incremental learning ability, while being able to reject unknown samples that are drawn from different distributions of the learned classes. This capability is crucial to the safety of deploying CIL models in open worlds. However, despite remarkable advancements in the respective CIL and OOD detection, there lacks a systematic and large-scale benchmark to assess the capability of advanced CIL models in detecting OOD samples. To fill this gap, in this study we design a comprehensive empirical study to establish such a benchmark, named $\textbf{OpenCIL}$. To this end, we propose two principled frameworks for enabling four representative CIL models with 15 diverse OOD detection methods, resulting in 60 baseline models for OOD detection in CIL. The empirical evaluation is performed on two popular CIL datasets with six commonly-used OOD datasets. One key observation we find through our comprehensive evaluation is that the CIL models can be severely biased towards the OOD samples and newly added classes when they are exposed to open environments. Motivated by this, we further propose a new baseline for OOD detection in CIL, namely Bi-directional Energy Regularization ($\textbf{BER}$), which is specially designed to mitigate these two biases in different CIL models by having energy regularization on both old and new classes. Its superior performance is justified in our experiments. All codes and datasets are open-source at $https://github.com/mala-lab/OpenCIL$.
Authors: Aray Karjauv, Sahin Albayrak
Abstract: Convolutional Neural Networks (CNNs) are known for their ability to learn hierarchical structures, naturally developing detectors for objects, and semantic concepts within their deeper layers. Activation maps (AMs) reveal these saliency regions, which are crucial for many Explainable AI (XAI) methods. However, the direct exploitation of raw AMs in CNNs for feature attribution remains underexplored in literature. This work revises Class Activation Map (CAM) methods by introducing the Label-free Activation Map (LaFAM), a streamlined approach utilizing raw AMs for feature attribution without reliance on labels. LaFAM presents an efficient alternative to conventional CAM methods, demonstrating particular effectiveness in saliency map generation for self-supervised learning while maintaining applicability in supervised learning scenarios.
Authors: Thomas Fel, Louis Bethune, Andrew Kyle Lampinen, Thomas Serre, Katherine Hermann
Abstract: Recent studies suggest that deep learning models inductive bias towards favoring simpler features may be one of the sources of shortcut learning. Yet, there has been limited focus on understanding the complexity of the myriad features that models learn. In this work, we introduce a new metric for quantifying feature complexity, based on $\mathscr{V}$-information and capturing whether a feature requires complex computational transformations to be extracted. Using this $\mathscr{V}$-information metric, we analyze the complexities of 10,000 features, represented as directions in the penultimate layer, that were extracted from a standard ImageNet-trained vision model. Our study addresses four key questions: First, we ask what features look like as a function of complexity and find a spectrum of simple to complex features present within the model. Second, we ask when features are learned during training. We find that simpler features dominate early in training, and more complex features emerge gradually. Third, we investigate where within the network simple and complex features flow, and find that simpler features tend to bypass the visual hierarchy via residual connections. Fourth, we explore the connection between features complexity and their importance in driving the networks decision. We find that complex features tend to be less important. Surprisingly, important features become accessible at earlier layers during training, like a sedimentation process, allowing the model to build upon these foundational elements.
Authors: Emaad Khwaja, Abdullah Rashwan, Ting Chen, Oliver Wang, Suraj Kothawade, Yeqing Li
Abstract: We present a one-shot text-to-image diffusion model that can generate high-resolution images from natural language descriptions. Our model employs a layered U-Net architecture that simultaneously synthesizes images at multiple resolution scales. We show that this method outperforms the baseline of synthesizing images only at the target resolution, while reducing the computational cost per step. We demonstrate that higher resolution synthesis can be achieved by layering convolutions at additional resolution scales, in contrast to other methods which require additional models for super-resolution synthesis.
Authors: Dejie Yang, Zhu Xu, Wentao Mo, Qingchao Chen, Siyuan Huang, Yang Liu
Abstract: 3D Vision-Language Pre-training (3D-VLP) aims to provide a pre-train model which can bridge 3D scenes with natural language, which is an important technique for embodied intelligence. However, current 3D-VLP datasets are hindered by limited scene-level diversity and insufficient fine-grained annotations (only 1.2K scenes and 280K textual annotations in ScanScribe), primarily due to the labor-intensive of collecting and annotating 3D scenes. To overcome these obstacles, we construct SynVL3D, a comprehensive synthetic scene-text corpus with 10K indoor scenes and 1M descriptions at object, view, and room levels, which has the advantages of diverse scene data, rich textual descriptions, multi-grained 3D-text associations, and low collection cost. Utilizing the rich annotations in SynVL3D, we pre-train a simple and unified Transformer for aligning 3D and language with multi-grained pretraining tasks. Moreover, we propose a synthetic-to-real domain adaptation in downstream task fine-tuning process to address the domain shift. Through extensive experiments, we verify the effectiveness of our model design by achieving state-of-the-art performance on downstream tasks including visual grounding, dense captioning, and question answering.
Authors: Nikhil Deekonda
Abstract: This paper introduces DogHeart, a dataset comprising 1400 training, 200 validation, and 400 test images categorized as small, normal, and large based on VHS score. A custom CNN model is developed, featuring a straightforward architecture with 4 convolutional layers and 4 fully connected layers. Despite the absence of data augmentation, the model achieves a 72\% accuracy in classifying cardiomegaly severity. The study contributes to automated assessment of cardiac conditions in dogs, highlighting the potential for early detection and intervention in veterinary care.
Authors: Xinyu Bai, Feng Xu
Abstract: Synthetic Aperture Radar (SAR) provides all-weather, high-resolution imaging capabilities, but its unique imaging mechanism often requires expert interpretation, limiting its widespread applicability. Translating SAR images into more easily recognizable optical images using diffusion models helps address this challenge. However, diffusion models suffer from high latency due to numerous iterative inferences, while Generative Adversarial Networks (GANs) can achieve image translation with just a single iteration but often at the cost of image quality. To overcome these issues, we propose a new training framework for SAR-to-optical image translation that combines the strengths of both approaches. Our method employs consistency distillation to reduce iterative inference steps and integrates adversarial learning to ensure image clarity and minimize color shifts. Additionally, our approach allows for a trade-off between quality and speed, providing flexibility based on application requirements. We conducted experiments on SEN12 and GF3 datasets, performing quantitative evaluations using Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Frechet Inception Distance (FID), as well as calculating the inference latency. The results demonstrate that our approach significantly improves inference speed by 131 times while maintaining the visual quality of the generated images, thus offering a robust and efficient solution for SAR-to-optical image translation.
Authors: Hasan Zohirul Islam, Safayet Khan, Sanjib Kumar Paul, Sheikh Imtiaz Rahi, Fahim Hossain Sifat, Md. Mahadi Hasan Sany, Md. Shahjahan Ali Sarker, Tareq Anam, Ismail Hossain Polas
Abstract: Absence of tamper-proof cattle identification technology was a significant problem preventing insurance companies from providing livestock insurance. This lack of technology had devastating financial consequences for marginal farmers as they did not have the opportunity to claim compensation for any unexpected events such as the accidental death of cattle in Bangladesh. Using machine learning and deep learning algorithms, we have solved the bottleneck of cattle identification by developing and introducing a muzzle-based cattle identification system. The uniqueness of cattle muzzles has been scientifically established, which resembles human fingerprints. This is the fundamental premise that prompted us to develop a cattle identification system that extracts the uniqueness of cattle muzzles. For this purpose, we collected 32,374 images from 826 cattle. Contrast-limited adaptive histogram equalization (CLAHE) with sharpening filters was applied in the preprocessing steps to remove noise from images. We used the YOLO algorithm for cattle muzzle detection in the image and the FaceNet architecture to learn unified embeddings from muzzle images using squared $L_2$ distances. Our system performs with an accuracy of $96.489\%$, $F_1$ score of $97.334\%$, and a true positive rate (tpr) of $87.993\%$ at a remarkably low false positive rate (fpr) of $0.098\%$. This reliable and efficient system for identifying cattle can significantly advance livestock insurance and precision farming.
Authors: Matt Pharr, Bartlomiej Wronski, Marco Salvi, Marcos Fajardo
Abstract: 2D texture maps and 3D voxel arrays are widely used to add rich detail to the surfaces and volumes of rendered scenes, and filtered texture lookups are integral to producing high-quality imagery. We show that applying the texture filter after evaluating shading generally gives more accurate imagery than filtering textures before BSDF evaluation, as is current practice. These benefits are not merely theoretical, but are apparent in common cases. We demonstrate that practical and efficient filtering after shading is possible through the use of stochastic sampling of texture filters. Stochastic texture filtering offers additional benefits, including efficient implementation of high-quality texture filters and efficient filtering of textures stored in compressed and sparse data structures, including neural representations. We demonstrate applications in both real-time and offline rendering and show that the additional error from stochastic filtering is minimal. We find that this error is handled well by either spatiotemporal denoising or moderate pixel sampling rates.
Authors: Jinhua Zhang, Hualian Sheng, Sijia Cai, Bing Deng, Qiao Liang, Wen Li, Ying Fu, Jieping Ye, Shuhang Gu
Abstract: Controllable generation is considered a potentially vital approach to address the challenge of annotating 3D data, and the precision of such controllable generation becomes particularly imperative in the context of data production for autonomous driving. Existing methods focus on the integration of diverse generative information into controlling inputs, utilizing frameworks such as GLIGEN or ControlNet, to produce commendable outcomes in controllable generation. However, such approaches intrinsically restrict generation performance to the learning capacities of predefined network architectures. In this paper, we explore the integration of controlling information and introduce PerlDiff (Perspective-Layout Diffusion Models), a method for effective street view image generation that fully leverages perspective 3D geometric information. Our PerlDiff employs 3D geometric priors to guide the generation of street view images with precise object-level control within the network learning process, resulting in a more robust and controllable output. Moreover, it demonstrates superior controllability compared to alternative layout control methods. Empirical results justify that our PerlDiff markedly enhances the precision of generation on the NuScenes and KITTI datasets. Our codes and models are publicly available at https://github.com/LabShuHangGU/PerlDiff.
Authors: Yashwardhan Chaudhuri, Ankit Kumar, Arun Balaji Buduru, Adel Alshamrani
Abstract: Crowd counting is gaining societal relevance, particularly in domains of Urban Planning, Crowd Management, and Public Safety. This paper introduces Fourier-guided attention (FGA), a novel attention mechanism for crowd count estimation designed to address the inefficient full-scale global pattern capture in existing works on convolution-based attention networks. FGA efficiently captures multi-scale information, including full-scale global patterns, by utilizing Fast-Fourier Transformations (FFT) along with spatial attention for global features and convolutions with channel-wise attention for semi-global and local features. The architecture of FGA involves a dual-path approach: (1) a path for processing full-scale global features through FFT, allowing for efficient extraction of information in the frequency domain, and (2) a path for processing remaining feature maps for semi-global and local features using traditional convolutions and channel-wise attention. This dual-path architecture enables FGA to seamlessly integrate frequency and spatial information, enhancing its ability to capture diverse crowd patterns. We apply FGA in the last layers of two popular crowd-counting works, CSRNet and CANNet, to evaluate the module's performance on benchmark datasets such as ShanghaiTech-A, ShanghaiTech-B, UCF-CC-50, and JHU++ crowd. The experiments demonstrate a notable improvement across all datasets based on Mean-Squared-Error (MSE) and Mean-Absolute-Error (MAE) metrics, showing comparable performance to recent state-of-the-art methods. Additionally, we illustrate the interpretability using qualitative analysis, leveraging Grad-CAM heatmaps, to show the effectiveness of FGA in capturing crowd patterns.
Authors: Rongchang Li, Zhenhua Feng, Tianyang Xu, Linze Li, Xiao-Jun Wu, Muhammad Awais, Sara Atito, Josef Kittler
Abstract: Compositional actions consist of dynamic (verbs) and static (objects) concepts. Humans can easily recognize unseen compositions using the learned concepts. For machines, solving such a problem requires a model to recognize unseen actions composed of previously observed verbs and objects, thus requiring, so-called, compositional generalization ability. To facilitate this research, we propose a novel Zero-Shot Compositional Action Recognition (ZS-CAR) task. For evaluating the task, we construct a new benchmark, Something-composition (Sth-com), based on the widely used Something-Something V2 dataset. We also propose a novel Component-to-Composition (C2C) learning method to solve the new ZS-CAR task. C2C includes an independent component learning module and a composition inference module. Last, we devise an enhanced training strategy to address the challenges of component variation between seen and unseen compositions and to handle the subtle balance between learning seen and unseen actions. The experimental results demonstrate that the proposed framework significantly surpasses the existing compositional generalization methods and sets a new state-of-the-art. The new Sth-com benchmark and code are available at https://github.com/RongchangLi/ZSCAR_C2C.
Authors: Nicholas Milef, John Keyser, Shu Kong
Abstract: Optical motion capture (mocap) requires accurately reconstructing the human body from retroreflective markers, including pose and shape. In a typical mocap setting, marker labeling is an important but tedious and error-prone step. Previous work has shown that marker labeling can be automated by using a structured template defining specific marker placements, but this places additional recording constraints. We propose to relax these constraints and solve for Unstructured Unlabeled Optical (UUO) mocap. Compared to the typical mocap setting that either labels markers or places them w.r.t a structured layout, markers in UUO mocap can be placed anywhere on the body and even on one specific limb (e.g., right leg for biomechanics research), hence it is of more practical significance. It is also more challenging. To solve UUO mocap, we exploit a monocular video captured by a single RGB camera, which does not require camera calibration. On this video, we run an off-the-shelf method to reconstruct and track a human individual, giving strong visual priors of human body pose and shape. With both the video and UUO markers, we propose an optimization pipeline towards marker identification, marker labeling, human pose estimation, and human body reconstruction. Our technical novelties include multiple hypothesis testing to optimize global orientation, and marker localization and marker-part matching to better optimize for body surface. We conduct extensive experiments to quantitatively compare our method against state-of-the-art approaches, including marker-only mocap and video-only human body/shape reconstruction. Experiments demonstrate that our method resoundingly outperforms existing methods on three established benchmark datasets for both full-body and partial-body reconstruction.
Authors: Qi Jia, Baoyu Fan, Cong Xu, Lu Liu, Liang Jin, Guoguang Du, Zhenhua Guo, Yaqian Zhao, Xuanjing Huang, Rengang Li
Abstract: Existing video multi-modal sentiment analysis mainly focuses on the sentiment expression of people within the video, yet often neglects the induced sentiment of viewers while watching the videos. Induced sentiment of viewers is essential for inferring the public response to videos, has broad application in analyzing public societal sentiment, effectiveness of advertising and other areas. The micro videos and the related comments provide a rich application scenario for viewers induced sentiment analysis. In light of this, we introduces a novel research task, Multi-modal Sentiment Analysis for Comment Response of Video Induced(MSA-CRVI), aims to inferring opinions and emotions according to the comments response to micro video. Meanwhile, we manually annotate a dataset named Comment Sentiment toward to Micro Video (CSMV) to support this research. It is the largest video multi-modal sentiment dataset in terms of scale and video duration to our knowledge, containing 107,267 comments and 8,210 micro videos with a video duration of 68.83 hours. To infer the induced sentiment of comment should leverage the video content, so we propose the Video Content-aware Comment Sentiment Analysis (VC-CSA) method as baseline to address the challenges inherent in this new task. Extensive experiments demonstrate that our method is showing significant improvements over other established baselines.
Authors: Zile Huang, Chong Zhang, Mingyu Jin, Fangyu Wu, Chengzhi Liu, Xiaobo Jin
Abstract: While deep learning-based general object detection has made significant strides in recent years, the effectiveness and efficiency of small object detection remain unsatisfactory. This is primarily attributed not only to the limited characteristics of such small targets but also to the high density and mutual overlap among these targets. The existing transformer-based small object detectors do not leverage the gap between accuracy and inference speed. To address challenges, we propose methods enhancing sampling within an end-to-end framework. Sample Points Refinement (SPR) constrains localization and attention, preserving meaningful interactions in the region of interest and filtering out misleading information. Scale-aligned Target (ST) integrates scale information into target confidence, improving classification for small object detection. A task-decoupled Sample Reweighting (SR) mechanism guides attention toward challenging positive examples, utilizing a weight generator module to assess the difficulty and adjust classification loss based on decoder layer outcomes. Comprehensive experiments across various benchmarks reveal that our proposed detector excels in detecting small objects. Our model demonstrates a significant enhancement, achieving a 2.9\% increase in average precision (AP) over the state-of-the-art (SOTA) on the VisDrone dataset and a 1.7\% improvement on the SODA-D dataset.
Authors: Guibin Zhao, Pengfei Li, Zhibo Zhang, Fusen Guo, Xueting Huang, Wei Xu, Jinyin Wang, Jianlong Chen
Abstract: Synthetic Aperture Radar has been extensively used in numerous fields and can gather a wealth of information about the area of interest. This large scene data intensive technology puts a high value on automatic target recognition which can free the utilizers and boost the efficiency. Recent advances in artificial intelligence have made it possible to create a deep learning based SAR ATR that can automatically identify target features from massive input data. In the last 6 years, intensive research has been conducted in this area, however, most papers in the current SAR ATR field used recurrent neural network and convolutional neural network varied models to deepen the regime's understanding of the SAR images. To equip SAR ATR with updated deep learning technology, this paper tries to apply a lightweight vision transformer based model to classify SAR images. The entire structure was verified by an open-accessed SAR data set and recognition results show that the final classification outcomes are robust and more accurate in comparison with referred traditional network structures without even using any convolutional layers.
Authors: Elaf A. Saeed, Ammar D. Jasim, Munther A. Abdul Malik
Abstract: Cuneiform writing, an old art style, allows us to see into the past. Aside from Egyptian hieroglyphs, the cuneiform script is one of the oldest writing systems. Many historians place Hebrew's origins in antiquity. For example, we used the same approach to decipher the cuneiform languages; after learning how to decipher one old language, we would visit an archaeologist to learn how to decipher any other ancient language. We propose a deep-learning-based sign detector method to speed up this procedure to identify and group cuneiform tablet images according to Hebrew letter content. The Hebrew alphabet is notoriously difficult and costly to gather the training data needed for deep learning, which entails enclosing Hebrew characters in boxes. We solve this problem using pre-existing transliterations and a sign-by-sign representation of the tablet's content in Latin characters. We recommend one of the supervised approaches because these do not include sign localization: We Find the transliteration signs in the tablet photographs by comparing them to their corresponding transliterations. Then, retrain the sign detector using these localized signs instead of utilizing annotations. Afterward, a more effective sign detector enhances the alignment quality. Consequently, this research aims to use the Yolov8 object identification pretraining model to identify Hebrew characters and categorize the cuneiform tablets.
Authors: Xiaojie Li, Yibo Yang, Jianlong Wu, Bernard Ghanem, Liqiang Nie, Min Zhang
Abstract: Few-shot class-incremental learning (FSCIL) confronts the challenge of integrating new classes into a model with minimal training samples while preserving the knowledge of previously learned classes. Traditional methods widely adopt static adaptation relying on a fixed parameter space to learn from data that arrive sequentially, prone to overfitting to the current session. Existing dynamic strategies require the expansion of the parameter space continually, leading to increased complexity. To address these challenges, we integrate the recently proposed selective state space model (SSM) into FSCIL. Concretely, we propose a dual selective SSM projector that dynamically adjusts the projection parameters based on the intermediate features for dynamic adaptation. The dual design enables the model to maintain the robust features of base classes, while adaptively learning distinctive feature shifts for novel classes. Additionally, we develop a class-sensitive selective scan mechanism to guide dynamic adaptation. It minimizes the disruption to base-class representations caused by training on novel data, and meanwhile, forces the selective scan to perform in distinct patterns between base and novel classes. Experiments on miniImageNet, CUB-200, and CIFAR-100 demonstrate that our framework outperforms the existing state-of-the-art methods. The code is available at https://github.com/xiaojieli0903/Mamba-FSCIL.
Authors: Manuel Lage Ca\~nellas, Le Nguyen, Anirban Mukherjee, Constantino \'Alvarez Casado, Xiaoting Wu, Praneeth Susarla, Sasan Sharifipour, Dinesh B. Jayagopi, Miguel Bordallo L\'opez
Abstract: In the domain of non-contact biometrics and human activity recognition, the lack of a versatile, multimodal dataset poses a significant bottleneck. To address this, we introduce the Oulu Multi Sensing (OMuSense-23) dataset that includes biosignals obtained from a mmWave radar, and an RGB-D camera. The dataset features data from 50 individuals in three distinct poses -- standing, sitting, and lying down -- each featuring four specific breathing pattern activities: regular breathing, reading, guided breathing, and apnea, encompassing both typical situations (e.g., sitting with normal breathing) and critical conditions (e.g., lying down without breathing). In our work, we present a detailed overview of the OMuSense-23 dataset, detailing the data acquisition protocol, describing the process for each participant. In addition, we provide, a baseline evaluation of several data analysis tasks related to biometrics, breathing pattern recognition and pose identification. Our results achieve a pose identification accuracy of 87\% and breathing pattern activity recognition of 83\% using features extracted from biosignals. The OMuSense-23 dataset is publicly available as resource for other researchers and practitioners in the field.
Authors: Harry Zhang, Luca Carlone
Abstract: We introduce CHAMP, a novel method for learning sequence-to-sequence, multi-hypothesis 3D human poses from 2D keypoints by leveraging a conditional distribution with a diffusion model. To predict a single output 3D pose sequence, we generate and aggregate multiple 3D pose hypotheses. For better aggregation results, we develop a method to score these hypotheses during training, effectively integrating conformal prediction into the learning process. This process results in a differentiable conformal predictor that is trained end2end with the 3D pose estimator. Post-training, the learned scoring model is used as the conformity score, and the 3D pose estimator is combined with a conformal predictor to select the most accurate hypotheses for downstream aggregation. Our results indicate that using a simple mean aggregation on the conformal prediction-filtered hypotheses set yields competitive results. When integrated with more sophisticated aggregation techniques, our method achieves state-of-the-art performance across various metrics and datasets while inheriting the probabilistic guarantees of conformal prediction.
Authors: Mohammad Reza Karimi Dastjerdi, Fr\'ed\'eric Fortier-Chouinard, Yannick Hold-Geoffroy, Marc H\'ebert, Claude Demers, Nima Kalantari, Jean-Fran\c{c}ois Lalonde
Abstract: Most novel view synthesis methods such as NeRF are unable to capture the true high dynamic range (HDR) radiance of scenes since they are typically trained on photos captured with standard low dynamic range (LDR) cameras. While the traditional exposure bracketing approach which captures several images at different exposures has recently been adapted to the multi-view case, we find such methods to fall short of capturing the full dynamic range of indoor scenes, which includes very bright light sources. In this paper, we present PanDORA: a PANoramic Dual-Observer Radiance Acquisition system for the casual capture of indoor scenes in high dynamic range. Our proposed system comprises two 360{\deg} cameras rigidly attached to a portable tripod. The cameras simultaneously acquire two 360{\deg} videos: one at a regular exposure and the other at a very fast exposure, allowing a user to simply wave the apparatus casually around the scene in a matter of minutes. The resulting images are fed to a NeRF-based algorithm that reconstructs the scene's full high dynamic range. Compared to HDR baselines from previous work, our approach reconstructs the full HDR radiance of indoor scenes without sacrificing the visual quality while retaining the ease of capture from recent NeRF-like approaches.
Authors: Young Chol Song
Abstract: Temporal grounding of activities, the identification of specific time intervals of actions within a larger event context, is a critical task in video understanding. Recent advancements in multimodal large language models (LLMs) offer new opportunities for enhancing temporal reasoning capabilities. In this paper, we evaluate the effectiveness of combining image-based and text-based large language models (LLMs) in a two-stage approach for temporal activity localization. We demonstrate that our method outperforms existing video-based LLMs. Furthermore, we explore the impact of instruction-tuning on a smaller multimodal LLM, showing that refining its ability to process action queries leads to more expressive and informative outputs, thereby enhancing its performance in identifying specific time intervals of activities. Our experimental results on the Charades-STA dataset highlight the potential of this approach in advancing the field of temporal activity localization and video understanding.
Authors: Xiaoli Zhang, Liying Wang, Libo Zhao, Xiongfei Li, Siwei Ma
Abstract: Multi-modality image fusion aims at fusing specific-modality and shared-modality information from two source images. To tackle the problem of insufficient feature extraction and lack of semantic awareness for complex scenes, this paper focuses on how to model correlation-driven decomposing features and reason high-level graph representation by efficiently extracting complementary features and multi-guided feature aggregation. We propose a three-branch encoder-decoder architecture along with corresponding fusion layers as the fusion strategy. The transformer with Multi-Dconv Transposed Attention and Local-enhanced Feed Forward network is used to extract shallow features after the depthwise convolution. In the three parallel branches encoder, Cross Attention and Invertible Block (CAI) enables to extract local features and preserve high-frequency texture details. Base feature extraction module (BFE) with residual connections can capture long-range dependency and enhance shared-modality expression capabilities. Graph Reasoning Module (GR) is introduced to reason high-level cross-modality relations and extract low-level details features as CAI's specific-modality complementary information simultaneously. Experiments demonstrate that our method has obtained competitive results compared with state-of-the-art methods in visible/infrared image fusion and medical image fusion tasks. Moreover, we surpass other fusion methods in terms of subsequent tasks, averagely scoring 9.78% mAP@.5 higher in object detection and 6.46% mIoU higher in semantic segmentation.
Authors: Khaled Alomar, Halil Ibrahim Aysel, Xiaohao Cai
Abstract: Human Action Recognition (HAR) encompasses the task of monitoring human activities across various domains, including but not limited to medical, educational, entertainment, visual surveillance, video retrieval, and the identification of anomalous activities. Over the past decade, the field of HAR has witnessed substantial progress by leveraging Convolutional Neural Networks (CNNs) to effectively extract and comprehend intricate information, thereby enhancing the overall performance of HAR systems. Recently, the domain of computer vision has witnessed the emergence of Vision Transformers (ViTs) as a potent solution. The efficacy of transformer architecture has been validated beyond the confines of image analysis, extending their applicability to diverse video-related tasks. Notably, within this landscape, the research community has shown keen interest in HAR, acknowledging its manifold utility and widespread adoption across various domains. This article aims to present an encompassing survey that focuses on CNNs and the evolution of Recurrent Neural Networks (RNNs) to ViTs given their importance in the domain of HAR. By conducting a thorough examination of existing literature and exploring emerging trends, this study undertakes a critical analysis and synthesis of the accumulated knowledge in this field. Additionally, it investigates the ongoing efforts to develop hybrid approaches. Following this direction, this article presents a novel hybrid model that seeks to integrate the inherent strengths of CNNs and ViTs.
Authors: Taiga Hayami, Hiroshi Watanabe
Abstract: Video compression technology is essential for transmitting and storing videos. Many video compression methods reduce information in videos by removing high-frequency components and utilizing similarities between frames. Alternatively, the implicit neural representations (INRs) for videos, which use networks to represent and compress videos through model compression. A conventional method improves the quality of reconstruction by using frame features. However, the detailed representation of the frames can be improved. To improve the quality of reconstructed frames, we propose a method that uses low-resolution frames as residual connection that is considered effective for image reconstruction. Experimental results show that our method outperforms the existing method, HNeRV, in PSNR for 46 of the 49 videos.
Authors: M. Rempe, F. H\"orst, C. Seibold, B. Hadaschik, M. Schlimbach, J. Egger, K. Kr\"oninger, F. Breuer, M. Blaimer, J. Kleesiek
Abstract: We present a novel preprocessing and prediction pipeline for the classification of magnetic resonance imaging (MRI) that takes advantage of the information rich complex valued k-Space. Using a publicly available MRI raw dataset with 312 subject and a total of 9508 slices, we show the advantage of utilizing the k-Space for better prostate cancer likelihood estimation in comparison to just using the magnitudinal information in the image domain, with an AUROC of $86.1\%\pm1.8\%$. Additionally, by using high undersampling rates and a simple principal component analysis (PCA) for coil compression, we reduce the time needed for reconstruction by avoiding the time intensive GRAPPA reconstruction algorithm. By using digital undersampling for our experiments, we show that scanning and reconstruction time could be reduced. Even with an undersampling factor of 16, our approach achieves meaningful results, with an AUROC of $71.4\%\pm2.9\%$, using the PCA coil combination and taking into account the k-Space information. With this study, we were able to show the feasibility of preserving phase and k-Space information, with consistent results. Besides preserving valuable information for further diagnostics, this approach can work without the time intensive ADC and reconstruction calculations, greatly reducing the post processing, as well as potential scanning time, increasing patient comfort and allowing a close to real-time prediction.
Authors: Aditya Annavajjala, Alind Khare, Animesh Agrawal, Igor Fedorov, Hugo Latapie, Myungjin Lee, Alexey Tumanov
Abstract: CNNs are increasingly deployed across different hardware, dynamic environments, and low-power embedded devices. This has led to the design and training of CNN architectures with the goal of maximizing accuracy subject to such variable deployment constraints. As the number of deployment scenarios grows, there is a need to find scalable solutions to design and train specialized CNNs. Once-for-all training has emerged as a scalable approach that jointly co-trains many models (subnets) at once with a constant training cost and finds specialized CNNs later. The scalability is achieved by training the full model and simultaneously reducing it to smaller subnets that share model weights (weight-shared shrinking). However, existing once-for-all training approaches incur huge training costs reaching 1200 GPU hours. We argue this is because they either start the process of shrinking the full model too early or too late. Hence, we propose Delayed $\epsilon$-Shrinking (D$\epsilon$pS) that starts the process of shrinking the full model when it is partially trained (~50%) which leads to training cost improvement and better in-place knowledge distillation to smaller models. The proposed approach also consists of novel heuristics that dynamically adjust subnet learning rates incrementally (E), leading to improved weight-shared knowledge distillation from larger to smaller subnets as well. As a result, DEpS outperforms state-of-the-art once-for-all training techniques across different datasets including CIFAR10/100, ImageNet-100, and ImageNet-1k on accuracy and cost. It achieves 1.83% higher ImageNet-1k top1 accuracy or the same accuracy with 1.3x reduction in FLOPs and 2.5x drop in training cost (GPU*hrs)
Authors: Julien Posso, Guy Bois, Yvon Savaria
Abstract: This article presents a pioneering approach to real-time spacecraft pose estimation, utilizing a mixed-precision quantized neural network implemented on the FPGA components of a commercially available Xilinx MPSoC, renowned for its suitability in space applications. Our co-design methodology includes a novel evaluation technique for assessing the layer-wise neural network sensitivity to quantization, facilitating an optimal balance between accuracy, latency, and FPGA resource utilization. Utilizing the FINN library, we developed a bespoke FPGA dataflow accelerator that integrates on-chip weights and activation functions to minimize latency and energy consumption. Our implementation is 7.7 times faster and 19.5 times more energy-efficient than the best-reported values in the existing spacecraft pose estimation literature. Furthermore, our contribution includes the first real-time, open-source implementation of such algorithms, marking a significant advancement in making efficient spacecraft pose estimation algorithms widely accessible. The source code is available at https://github.com/possoj/FPGA-SpacePose.
Authors: Hannah Lee, Changyeon Lee, Kevin Farhat, Lin Qiu, Steve Geluso, Aerin Kim, Oren Etzioni
Abstract: Multimodal generative models are rapidly evolving, leading to a surge in the generation of realistic video and audio that offers exciting possibilities but also serious risks. Deepfake videos, which can convincingly impersonate individuals, have particularly garnered attention due to their potential misuse in spreading misinformation and creating fraudulent content. This survey paper examines the dual landscape of deepfake video generation and detection, emphasizing the need for effective countermeasures against potential abuses. We provide a comprehensive overview of current deepfake generation techniques, including face swapping, reenactment, and audio-driven animation, which leverage cutting-edge technologies like generative adversarial networks and diffusion models to produce highly realistic fake videos. Additionally, we analyze various detection approaches designed to differentiate authentic from altered videos, from detecting visual artifacts to deploying advanced algorithms that pinpoint inconsistencies across video and audio signals. The effectiveness of these detection methods heavily relies on the diversity and quality of datasets used for training and evaluation. We discuss the evolution of deepfake datasets, highlighting the importance of robust, diverse, and frequently updated collections to enhance the detection accuracy and generalizability. As deepfakes become increasingly indistinguishable from authentic content, developing advanced detection techniques that can keep pace with generation technologies is crucial. We advocate for a proactive approach in the "tug-of-war" between deepfake creators and detectors, emphasizing the need for continuous research collaboration, standardization of evaluation metrics, and the creation of comprehensive benchmarks.
Authors: Zhhengyong Huang, Yao Sui
Abstract: Image segmentation is critically important in almost all medical image analysis for automatic interpretations and processing. However, it is often challenging to perform image segmentation due to data imbalance between intra- and inter-class, resulting in over- or under-segmentation. Consequently, we proposed a new methodology to address the above issue, with a compact yet effective contour-weighted loss function. Our new loss function incorporates a contour-weighted cross-entropy loss and separable dice loss. The former loss extracts the contour of target regions via morphological erosion and generates a weight map for the cross-entropy criterion, whereas the latter divides the target regions into contour and non-contour components through the extracted contour map, calculates dice loss separately, and combines them to update the network. We carried out abdominal organ segmentation and brain tumor segmentation on two public datasets to assess our approach. Experimental results demonstrated that our approach offered superior segmentation, as compared to several state-of-the-art methods, while in parallel improving the robustness of those popular state-of-the-art deep models through our new loss function. The code is available at https://github.com/huangzyong/Contour-weighted-Loss-Seg.
URLs: https://github.com/huangzyong/Contour-weighted-Loss-Seg.
Authors: Antonia Karamolegkou, Phillip Rust, Yong Cao, Ruixiang Cui, Anders S{\o}gaard, Daniel Hershcovich
Abstract: Large vision-language models (VLMs) can assist visually impaired people by describing images from their daily lives. Current evaluation datasets may not reflect diverse cultural user backgrounds or the situational context of this use case. To address this problem, we create a survey to determine caption preferences and propose a culture-centric evaluation benchmark by filtering VizWiz, an existing dataset with images taken by people who are blind. We then evaluate several VLMs, investigating their reliability as visual assistants in a culturally diverse setting. While our results for state-of-the-art models are promising, we identify challenges such as hallucination and misalignment of automatic evaluation metrics with human judgment. We make our survey, data, code, and model outputs publicly available.
Authors: Anthony Miyaguchi, Murilo Gustineli, Austin Fischer, Ryan Lundqvist
Abstract: We present our approach for the SnakeCLEF 2024 competition to predict snake species from images. We explore and use Meta's DINOv2 vision transformer model for feature extraction to tackle species' high variability and visual similarity in a dataset of 182,261 images. We perform exploratory analysis on embeddings to understand their structure, and train a linear classifier on the embeddings to predict species. Despite achieving a score of 39.69, our results show promise for DINOv2 embeddings in snake identification. All code for this project is available at https://github.com/dsgt-kaggle-clef/snakeclef-2024.
Authors: Xingyi Yang, Xinchao Wang
Abstract: Large-scale Text-to-Video (T2V) diffusion models have recently demonstrated unprecedented capability to transform natural language descriptions into stunning and photorealistic videos. Despite the promising results, a significant challenge remains: these models struggle to fully grasp complex compositional interactions between multiple concepts and actions. This issue arises when some words dominantly influence the final video, overshadowing other concepts.To tackle this problem, we introduce \textbf{Vico}, a generic framework for compositional video generation that explicitly ensures all concepts are represented properly. At its core, Vico analyzes how input tokens influence the generated video, and adjusts the model to prevent any single concept from dominating. Specifically, Vico extracts attention weights from all layers to build a spatial-temporal attention graph, and then estimates the influence as the \emph{max-flow} from the source text token to the video target token. Although the direct computation of attention flow in diffusion models is typically infeasible, we devise an efficient approximation based on subgraph flows and employ a fast and vectorized implementation, which in turn makes the flow computation manageable and differentiable. By updating the noisy latent to balance these flows, Vico captures complex interactions and consequently produces videos that closely adhere to textual descriptions. We apply our method to multiple diffusion-based video models for compositional T2V and video editing. Empirical results demonstrate that our framework significantly enhances the compositional richness and accuracy of the generated videos. Visit our website at~\href{https://adamdad.github.io/vico/}{\url{https://adamdad.github.io/vico/}}.
URLs: https://adamdad.github.io/vico/, https://adamdad.github.io/vico/
Authors: Yu Zeng, Vishal M. Patel, Haochen Wang, Xun Huang, Ting-Chun Wang, Ming-Yu Liu, Yogesh Balaji
Abstract: Personalized text-to-image generation models enable users to create images that depict their individual possessions in diverse scenes, finding applications in various domains. To achieve the personalization capability, existing methods rely on finetuning a text-to-image foundation model on a user's custom dataset, which can be non-trivial for general users, resource-intensive, and time-consuming. Despite attempts to develop finetuning-free methods, their generation quality is much lower compared to their finetuning counterparts. In this paper, we propose Joint-Image Diffusion (\jedi), an effective technique for learning a finetuning-free personalization model. Our key idea is to learn the joint distribution of multiple related text-image pairs that share a common subject. To facilitate learning, we propose a scalable synthetic dataset generation technique. Once trained, our model enables fast and easy personalization at test time by simply using reference images as input during the sampling process. Our approach does not require any expensive optimization process or additional modules and can faithfully preserve the identity represented by any number of reference images. Experimental results show that our model achieves state-of-the-art generation quality, both quantitatively and qualitatively, significantly outperforming both the prior finetuning-based and finetuning-free personalization baselines.
Authors: Xinying Guo, Mingyuan Zhang, Haozhe Xie, Chenyang Gu, Ziwei Liu
Abstract: Crowd Motion Generation is essential in entertainment industries such as animation and games as well as in strategic fields like urban simulation and planning. This new task requires an intricate integration of control and generation to realistically synthesize crowd dynamics under specific spatial and semantic constraints, whose challenges are yet to be fully explored. On the one hand, existing human motion generation models typically focus on individual behaviors, neglecting the complexities of collective behaviors. On the other hand, recent methods for multi-person motion generation depend heavily on pre-defined scenarios and are limited to a fixed, small number of inter-person interactions, thus hampering their practicality. To overcome these challenges, we introduce CrowdMoGen, a zero-shot text-driven framework that harnesses the power of Large Language Model (LLM) to incorporate the collective intelligence into the motion generation framework as guidance, thereby enabling generalizable planning and generation of crowd motions without paired training data. Our framework consists of two key components: 1) Crowd Scene Planner that learns to coordinate motions and dynamics according to specific scene contexts or introduced perturbations, and 2) Collective Motion Generator that efficiently synthesizes the required collective motions based on the holistic plans. Extensive quantitative and qualitative experiments have validated the effectiveness of our framework, which not only fills a critical gap by providing scalable and generalizable solutions for Crowd Motion Generation task but also achieves high levels of realism and flexibility.
Authors: Orr Zohar, Xiaohan Wang, Yonatan Bitton, Idan Szpektor, Serena Yeung-Levy
Abstract: The performance of Large Vision Language Models (LVLMs) is dependent on the size and quality of their training datasets. Existing video instruction tuning datasets lack diversity as they are derived by prompting large language models with video captions to generate question-answer pairs, and are therefore mostly descriptive. Meanwhile, many labeled video datasets with diverse labels and supervision exist - however, we find that their integration into LVLMs is non-trivial. Herein, we present Video Self-Training with augmented Reasoning (Video-STaR), the first video self-training approach. Video-STaR allows the utilization of any labeled video dataset for video instruction tuning. In Video-STaR, an LVLM cycles between instruction generation and finetuning, which we show (I) improves general video understanding and (II) adapts LVLMs to novel downstream tasks with existing supervision. During generation, an LVLM is prompted to propose an answer. The answers are then filtered only to those that contain the original video labels, and the LVLM is then re-trained on the generated dataset. By only training on generated answers that contain the correct video labels, Video-STaR utilizes these existing video labels as weak supervision for video instruction tuning. Our results demonstrate that Video-STaR-enhanced LVLMs exhibit improved performance in (I) general video QA, where TempCompass performance improved by 10%, and (II) on downstream tasks, where Video-STaR improved Kinetics700-QA accuracy by 20% and action quality assessment on FineDiving by 15%.
Authors: Xiang Xu, Lingdong Kong, Hui Shuai, Wenwei Zhang, Liang Pan, Kai Chen, Ziwei Liu, Qingshan Liu
Abstract: In the realm of autonomous driving, accurate 3D perception is the foundation. However, developing such models relies on extensive human annotations -- a process that is both costly and labor-intensive. To address this challenge from a data representation learning perspective, we introduce SuperFlow, a novel framework designed to harness consecutive LiDAR-camera pairs for establishing spatiotemporal pretraining objectives. SuperFlow stands out by integrating two key designs: 1) a dense-to-sparse consistency regularization, which promotes insensitivity to point cloud density variations during feature learning, and 2) a flow-based contrastive learning module, carefully crafted to extract meaningful temporal cues from readily available sensor calibrations. To further boost learning efficiency, we incorporate a plug-and-play view consistency module that enhances the alignment of the knowledge distilled from camera views. Extensive comparative and ablation studies across 11 heterogeneous LiDAR datasets validate our effectiveness and superiority. Additionally, we observe several interesting emerging properties by scaling up the 2D and 3D backbones during pretraining, shedding light on the future research of 3D foundation models for LiDAR-based perception.
Authors: Zhangyang Qi, Yunhan Yang, Mengchen Zhang, Long Xing, Xiaoyang Wu, Tong Wu, Dahua Lin, Xihui Liu, Jiaqi Wang, Hengshuang Zhao
Abstract: Recent advances in 3D AIGC have shown promise in directly creating 3D objects from text and images, offering significant cost savings in animation and product design. However, detailed edit and customization of 3D assets remains a long-standing challenge. Specifically, 3D Generation methods lack the ability to follow finely detailed instructions as precisely as their 2D image creation counterparts. Imagine you can get a toy through 3D AIGC but with undesired accessories and dressing. To tackle this challenge, we propose a novel pipeline called Tailor3D, which swiftly creates customized 3D assets from editable dual-side images. We aim to emulate a tailor's ability to locally change objects or perform overall style transfer. Unlike creating 3D assets from multiple views, using dual-side images eliminates conflicts on overlapping areas that occur when editing individual views. Specifically, it begins by editing the front view, then generates the back view of the object through multi-view diffusion. Afterward, it proceeds to edit the back views. Finally, a Dual-sided LRM is proposed to seamlessly stitch together the front and back 3D features, akin to a tailor sewing together the front and back of a garment. The Dual-sided LRM rectifies imperfect consistencies between the front and back views, enhancing editing capabilities and reducing memory burdens while seamlessly integrating them into a unified 3D representation with the LoRA Triplane Transformer. Experimental results demonstrate Tailor3D's effectiveness across various 3D generation and editing tasks, including 3D generative fill and style transfer. It provides a user-friendly, efficient solution for editing 3D assets, with each editing step taking only seconds to complete.
Authors: Xuweiyi Chen, Ziqiao Ma, Xuejun Zhang, Sihan Xu, Shengyi Qian, Jianing Yang, David F. Fouhey, Joyce Chai
Abstract: Large vision language models (LVLMs) often suffer from object hallucination, producing objects not present in the given images. While current benchmarks for object hallucination primarily concentrate on the presence of a single object class rather than individual entities, this work systematically investigates multi-object hallucination, examining how models misperceive (e.g., invent nonexistent objects or become distracted) when tasked with focusing on multiple objects simultaneously. We introduce Recognition-based Object Probing Evaluation (ROPE), an automated evaluation protocol that considers the distribution of object classes within a single image during testing and uses visual referring prompts to eliminate ambiguity. With comprehensive empirical studies and analysis of potential factors leading to multi-object hallucination, we found that (1) LVLMs suffer more hallucinations when focusing on multiple objects compared to a single object. (2) The tested object class distribution affects hallucination behaviors, indicating that LVLMs may follow shortcuts and spurious correlations.(3) Hallucinatory behaviors are influenced by data-specific factors, salience and frequency, and model intrinsic behaviors. We hope to enable LVLMs to recognize and reason about multiple objects that often occur in realistic visual scenes, provide insights, and quantify our progress towards mitigating the issues.
Authors: Min-Hyeok Sun, Dong-Hee Paek, Seung-Hyun Song, Seung-Hyun Kong
Abstract: Focusing on the strength of 4D (4-Dimensional) radar, research about robust 3D object detection networks in adverse weather conditions has gained attention. To train such networks, datasets that contain large amounts of 4D radar data and ground truth labels are essential. However, the existing 4D radar datasets (e.g., K-Radar) lack sufficient sensor data and labels, which hinders the advancement in this research domain. Furthermore, enlarging the 4D radar datasets requires a time-consuming and expensive manual labeling process. To address these issues, we propose the auto-labeling method of 4D radar tensor (4DRT) in the K-Radar dataset. The proposed method initially trains a LiDAR-based object detection network (LODN) using calibrated LiDAR point cloud (LPC). The trained LODN then automatically generates ground truth labels (i.e., auto-labels, ALs) of the K-Radar train dataset without human intervention. The generated ALs are used to train the 4D radar-based object detection network (4DRODN), Radar Tensor Network with Height (RTNH). The experimental results demonstrate that RTNH trained with ALs has achieved a similar detection performance to the original RTNH which is trained with manually annotated ground truth labels, thereby verifying the effectiveness of the proposed auto-labeling method. All relevant codes will be soon available at the following GitHub project: https://github.com/kaist-avelab/K-Radar
Authors: Iman Kianian, Hedieh Sajedi
Abstract: Brain age estimation involves predicting the biological age of individuals from their brain images, which offers valuable insights into the aging process and the progression of neurodegenerative diseases. Conducting large-scale datasets for medical image analysis is a challenging and time-consuming task. Existing approaches mostly depend on large datasets, which are hard to come by and expensive. These approaches also require sophisticated, resource-intensive models with a large number of parameters, necessitating a considerable amount of processing power. As a result, there is a vital need to develop innovative methods that can achieve robust performance with limited datasets and efficient use of computational resources. This paper proposes a novel slice-based dual-stream method called GDSM (Greedy Dual-Stream Model) for brain age estimation. This method addresses the limitations of large dataset requirements and computational resource intensiveness. The proposed method incorporates local and global aspects of the brain, thereby refining the focus on specific target regions. The approach employs four backbones to predict ages based on local and global features, complemented by a final model for age correction. Our method demonstrates a Mean Absolute Error (MAE) of 3.25 years on the test set of IBID, which only contains 289 subjects. To demonstrate the robustness of our approach for any small dataset, we analyzed the proposed method with the IXI dataset and achieved an MAE of 4.18 years on the test set of IXI. By leveraging dual-stream and greedy strategies, this approach achieves efficiency and robust performance, making it comparable with other state-of-the-art methods. The code for the GDSM model is available at https://github.com/iman2693/GDSM.
Authors: Christian Quaia, Richard J Krauzlis
Abstract: If neuroscientists were asked which brain area is responsible for object recognition in primates, most would probably answer infero-temporal (IT) cortex. While IT is likely responsible for fine discriminations, and it is accordingly dominated by foveal visual inputs, there is more to object recognition than fine discrimination. Importantly, foveation of an object of interest usually requires recognizing, with reasonable confidence, its presence in the periphery. Arguably, IT plays a secondary role in such peripheral recognition, and other visual areas might instead be more critical. To investigate how signals carried by early visual processing areas (such as LGN and V1) could be used for object recognition in the periphery, we focused here on the task of distinguishing faces from non-faces. We tested how sensitive various models were to nuisance parameters, such as changes in scale and orientation of the image, and the type of image background. We found that a model of V1 simple or complex cells could provide quite reliable information, resulting in performance better than 80% in realistic scenarios. An LGN model performed considerably worse. Because peripheral recognition is both crucial to enable fine recognition (by bringing an object of interest on the fovea), and probably sufficient to account for a considerable fraction of our daily recognition-guided behavior, we think that the current focus on area IT and foveal processing is too narrow. We propose that rather than a hierarchical system with IT-like properties as its primary aim, object recognition should be seen as a parallel process, with high-accuracy foveal modules operating in parallel with lower-accuracy and faster modules that can operate across the visual field.
Authors: Jiawei Zhang
Abstract: In this paper, we will introduce a novel deep model named Reconciled Polynomial Network (RPN) for deep function learning. RPN has a very general architecture and can be used to build models with various complexities, capacities, and levels of completeness, which all contribute to the correctness of these models. As indicated in the subtitle, RPN can also serve as the backbone to unify different base models into one canonical representation. This includes non-deep models, like probabilistic graphical models (PGMs) - such as Bayesian network and Markov network - and kernel support vector machines (kernel SVMs), as well as deep models like the classic multi-layer perceptron (MLP) and the recent Kolmogorov-Arnold network (KAN). Technically, RPN proposes to disentangle the underlying function to be inferred into the inner product of a data expansion function and a parameter reconciliation function. Together with the remainder function, RPN accurately approximates the underlying functions that governs data distributions. The data expansion functions in RPN project data vectors from the input space to a high-dimensional intermediate space, specified by the expansion functions in definition. Meanwhile, RPN also introduces the parameter reconciliation functions to fabricate a small number of parameters into a higher-order parameter matrix to address the ``curse of dimensionality'' problem caused by the data expansions. Moreover, the remainder functions provide RPN with additional complementary information to reduce potential approximation errors. We conducted extensive empirical experiments on numerous benchmark datasets across multiple modalities, including continuous function datasets, discrete vision and language datasets, and classic tabular datasets, to investigate the effectiveness of RPN.
Authors: Shirley Kokane, Mostofa Rafid Uddin, Min Xu
Abstract: Transfer learning methods start performing poorly when the complexity of the learning task is increased. Most of these methods calculate the cumulative differences of all the matched features and then use them to back-propagate that loss through all the layers. Contrary to these methods, in this work, we propose a novel layer-wise learning scheme that adjusts learning parameters per layer as a function of the differences in the Jacobian/Attention/Hessian of the output activations w.r.t. the network parameters. We applied this novel scheme for attention map-based and derivative-based (first and second order) transfer learning methods. We received improved learning performance and stability against a wide range of datasets. From extensive experimental evaluation, we observed that the performance boost achieved by our method becomes more significant with the increasing difficulty of the learning task.
Authors: Mahdi Ait Lhaj Loutfi, Teodora Boblea Podasca, Alex Zwanenburg, Taman Upadhaya, Jorge Barrios, David R. Raleigh, William C. Chen, Dante P. I. Capaldi, Hong Zheng, Olivier Gevaert, Jing Wu, Alvin C. Silva, Paul J. Zhang, Harrison X. Bai, Jan Seuntjens, Steffen L\"ock, Patrick O. Richard, Olivier Morin, Caroline Reinhold, Martin Lepage, Martin Valli\`eres
Abstract: Background: The high dimensionality of radiomic feature sets, the variability in radiomic feature types and potentially high computational requirements all underscore the need for an effective method to identify the smallest set of predictive features for a given clinical problem. Purpose: Develop a methodology and tools to identify and explain the smallest set of predictive radiomic features. Materials and Methods: 89,714 radiomic features were extracted from five cancer datasets: low-grade glioma, meningioma, non-small cell lung cancer (NSCLC), and two renal cell carcinoma cohorts (n=2104). Features were categorized by computational complexity into morphological, intensity, texture, linear filters, and nonlinear filters. Models were trained and evaluated on each complexity level using the area under the curve (AUC). The most informative features were identified, and their importance was explained. The optimal complexity level and associated most informative features were identified using systematic statistical significance analyses and a false discovery avoidance procedure, respectively. Their predictive importance was explained using a novel tree-based method. Results: MEDimage, a new open-source tool, was developed to facilitate radiomic studies. Morphological features were optimal for MRI-based meningioma (AUC: 0.65) and low-grade glioma (AUC: 0.68). Intensity features were optimal for CECT-based renal cell carcinoma (AUC: 0.82) and CT-based NSCLC (AUC: 0.76). Texture features were optimal for MRI-based renal cell carcinoma (AUC: 0.72). Tuning the Hounsfield unit range improved results for CECT-based renal cell carcinoma (AUC: 0.86). Conclusion: Our proposed methodology and software can estimate the optimal radiomics complexity level for specific medical outcomes, potentially simplifying the use of radiomics in predictive modeling across various contexts.
Authors: Zekun Li, Xianjun Yang, Kyuri Choi, Wanrong Zhu, Ryan Hsieh, HyeonJung Kim, Jin Hyuk Lim, Sungyoung Ji, Byungju Lee, Xifeng Yan, Linda Ruth Petzold, Stephen D. Wilson, Woosang Lim, William Yang Wang
Abstract: The rapid advancement of Large Language Models (LLMs) and Large Multimodal Models (LMMs) has heightened the demand for AI-based scientific assistants capable of understanding scientific articles and figures. Despite progress, there remains a significant gap in evaluating models' comprehension of professional, graduate-level, and even PhD-level scientific content. Current datasets and benchmarks primarily focus on relatively simple scientific tasks and figures, lacking comprehensive assessments across diverse advanced scientific disciplines. To bridge this gap, we collected a multimodal, multidisciplinary dataset from open-access scientific articles published in Nature Communications journals. This dataset spans 72 scientific disciplines, ensuring both diversity and quality. We created benchmarks with various tasks and settings to comprehensively evaluate LMMs' capabilities in understanding scientific figures and content. Our evaluation revealed that these tasks are highly challenging: many open-source models struggled significantly, and even GPT-4V and GPT-4o faced difficulties. We also explored using our dataset as training resources by constructing visual instruction-following data, enabling the 7B LLaVA model to achieve performance comparable to GPT-4V/o on our benchmark. Additionally, we investigated the use of our interleaved article texts and figure images for pre-training LMMs, resulting in improvements on the material generation task. The source dataset, including articles, figures, constructed benchmarks, and visual instruction-following data, is open-sourced.
Authors: Hang Chen, Sankepally Sainath Reddy, Ziwei Chen, Dianbo Liu
Abstract: The dimensionality of the embedding and the number of available embeddings ( also called codebook size) are critical factors influencing the performance of Vector Quantization(VQ), a discretization process used in many models such as the Vector Quantized Variational Autoencoder (VQ-VAE) architecture. This study examines the balance between the codebook sizes and dimensions of embeddings in VQ, while maintaining their product constant. Traditionally, these hyper parameters are static during training; however, our findings indicate that augmenting the codebook size while simultaneously reducing the embedding dimension can significantly boost the effectiveness of the VQ-VAE. As a result, the strategic selection of codebook size and embedding dimensions, while preserving the capacity of the discrete codebook space, is critically important. To address this, we propose a novel adaptive dynamic quantization approach, underpinned by the Gumbel-Softmax mechanism, which allows the model to autonomously determine the optimal codebook configuration for each data instance. This dynamic discretizer gives the VQ-VAE remarkable flexibility. Thorough empirical evaluations across multiple benchmark datasets validate the notable performance enhancements achieved by our approach, highlighting the significant potential of adaptive dynamic quantization to improve model performance.
Authors: Georgiy Kiselev
Abstract: This paper demonstrates the efficacy of a modified U-Net structure for the extraction of vascular tree masks for human fundus photographs. On limited compute resources and training data, the proposed model only slightly underperforms when compared to state of the art methods.
Authors: Ethan Mendes, Yang Chen, James Hays, Sauvik Das, Wei Xu, Alan Ritter
Abstract: Vision Language Models (VLMs) are rapidly advancing in their capability to answer information-seeking questions. As these models are widely deployed in consumer applications, they could lead to new privacy risks due to emergent abilities to identify people in photos, geolocate images, etc. As we demonstrate, somewhat surprisingly, current open-source and proprietary VLMs are very capable image geolocators, making widespread geolocation with VLMs an immediate privacy risk, rather than merely a theoretical future concern. As a first step to address this challenge, we develop a new benchmark, GPTGeoChat, to test the ability of VLMs to moderate geolocation dialogues with users. We collect a set of 1,000 image geolocation conversations between in-house annotators and GPT-4v, which are annotated with the granularity of location information revealed at each turn. Using this new dataset, we evaluate the ability of various VLMs to moderate GPT-4v geolocation conversations by determining when too much location information has been revealed. We find that custom fine-tuned models perform on par with prompted API-based models when identifying leaked location information at the country or city level; however, fine-tuning on supervised data appears to be needed to accurately moderate finer granularities, such as the name of a restaurant or building.
Authors: Sree Rama Vamsidhar S, Bhargava Satya, Rama Krishna Gorthi
Abstract: Deep Learning (DL) approaches have gained prominence in medical imaging for disease diagnosis. Chest X-ray (CXR) classification has emerged as an effective method for detecting various diseases. Among these methodologies, Chest X-ray (CXR) classification has proven to be an effective approach for detecting and analyzing various diseases. However, the reliable performance of DL classification algorithms is dependent upon access to large and balanced datasets, which pose challenges in medical imaging due to the impracticality of acquiring sufficient data for every disease category. To tackle this problem, we propose an algorithmic-centric approach called Effective-Label Distribution Aware Margin (E-LDAM), which modifies the margin of the widely adopted Label Distribution Aware Margin (LDAM) loss function using an effective number of samples in each class. Experimental evaluations on the COVIDx CXR dataset focus on Normal, Pneumonia, and COVID-19 classification. The experimental results demonstrate the effectiveness of the proposed E-LDAM approach, achieving a remarkable recall score of 97.81% for the minority class (COVID-19) in CXR image prediction. Furthermore, the overall accuracy of the three-class classification task attains an impressive level of 95.26%.
Authors: Wei Chen, Shian Du, Shigui Li, Delu Zeng, John Paisley
Abstract: Normalizing Flows (NFs) have gained popularity among deep generative models due to their ability to provide exact likelihood estimation and efficient sampling. However, a crucial limitation of NFs is their substantial memory requirements, arising from maintaining the dimension of the latent space equal to that of the input space. Multi-scale architectures bypass this limitation by progressively reducing the dimension of latent variables while ensuring reversibility. Existing multi-scale architectures split the latent variables in a simple, static manner at the channel level, compromising NFs' expressive power. To address this issue, we propose a regularized and feature-dependent $\mathtt{Shuffle}$ operation and integrate it into vanilla multi-scale architecture. This operation heuristically generates channel-wise weights and adaptively shuffles latent variables before splitting them with these weights. We observe that such operation guides the variables to evolve in the direction of entropy increase, hence we refer to NFs with the $\mathtt{Shuffle}$ operation as \emph{Entropy-Informed Weighting Channel Normalizing Flow} (EIW-Flow). Experimental results indicate that the EIW-Flow achieves state-of-the-art density estimation results and comparable sample quality on CIFAR-10, CelebA and ImageNet datasets, with negligible additional computational overhead.
Authors: Yijia Xiao, Edward Sun, Tianyu Liu, Wei Wang
Abstract: We propose LogicVista, an evaluation benchmark that assesses the integrated logical reasoning capabilities of multimodal large language models (MLLMs) in Visual contexts. Recent advancements in MLLMs have demonstrated various fascinating abilities, from crafting poetry based on an image to performing mathematical reasoning. However, there is still a lack of systematic evaluation of MLLMs' proficiency in logical reasoning tasks, which are essential for activities like navigation and puzzle-solving. Thus we evaluate general logical cognition abilities across 5 logical reasoning tasks encompassing 9 different capabilities, using a sample of 448 multiple-choice questions. Each question is annotated with the correct answer and the human-written reasoning behind the selection, enabling both open-ended and multiple-choice evaluation. A total of 8 MLLMs are comprehensively evaluated using LogicVista. Code and Data Available at https://github.com/Yijia-Xiao/LogicVista.
Authors: Sishun Pan, Xixian Wu, Tingmin Li, Longfei Huang, Mingxu Feng, Zhonghua Wan, Yang Yang
Abstract: This paper presents a data-free, parameter-isolation-based continual learning algorithm we developed for the sequential task continual learning track of the 2nd Greater Bay Area International Algorithm Competition. The method learns an independent parameter subspace for each task within the network's convolutional and linear layers and freezes the batch normalization layers after the first task. Specifically, for domain incremental setting where all domains share a classification head, we freeze the shared classification head after first task is completed, effectively solving the issue of catastrophic forgetting. Additionally, facing the challenge of domain incremental settings without providing a task identity, we designed an inference task identity strategy, selecting an appropriate mask matrix for each sample. Furthermore, we introduced a gradient supplementation strategy to enhance the importance of unselected parameters for the current task, facilitating learning for new tasks. We also implemented an adaptive importance scoring strategy that dynamically adjusts the amount of parameters to optimize single-task performance while reducing parameter usage. Moreover, considering the limitations of storage space and inference time, we designed a mask matrix compression strategy to save storage space and improve the speed of encryption and decryption of the mask matrix. Our approach does not require expanding the core network or using external auxiliary networks or data, and performs well under both task incremental and domain incremental settings. This solution ultimately won a second-place prize in the competition.
Authors: Kyobin Choo, Youngjun Jun, Mijin Yun, Seong Jae Hwang
Abstract: In neuroimaging, generally, brain CT is more cost-effective and accessible imaging option compared to MRI. Nevertheless, CT exhibits inferior soft-tissue contrast and higher noise levels, yielding less precise structural clarity. In response, leveraging more readily available CT to construct its counterpart MRI, namely, medical image-to-image translation (I2I), serves as a promising solution. Particularly, while diffusion models (DMs) have recently risen as a powerhouse, they also come with a few practical caveats for medical I2I. First, DMs' inherent stochasticity from random noise sampling cannot guarantee consistent MRI generation that faithfully reflects its CT. Second, for 3D volumetric images which are prevalent in medical imaging, naively using 2D DMs leads to slice inconsistency, e.g., abnormal structural and brightness changes. While 3D DMs do exist, significant training costs and data dependency bring hesitation. As a solution, we propose novel style key conditioning (SKC) and inter-slice trajectory alignment (ISTA) sampling for the 2D Brownian bridge diffusion model. Specifically, SKC ensures a consistent imaging style (e.g., contrast) across slices, and ISTA interconnects the independent sampling of each slice, deterministically achieving style and shape consistent 3D CT-to-MRI translation. To the best of our knowledge, this study is the first to achieve high-quality 3D medical I2I based only on a 2D DM with no extra architectural models. Our experimental results show superior 3D medical I2I than existing 2D and 3D baselines, using in-house CT-MRI dataset and BraTS2023 FLAIR-T1 MRI dataset.
Authors: Yuan Gao, Shuguo Jiang, Moran Li, Jin-Gang Yu, Gui-Song Xia
Abstract: We aim to address Multi-Task Learning (MTL) with a large number of tasks by Multi-Task Grouping (MTG). Given N tasks, we propose to simultaneously identify the best task groups from 2^N candidates and train the model weights simultaneously in one-shot, with the high-order task-affinity fully exploited. This is distinct from the pioneering methods which sequentially identify the groups and train the model weights, where the group identification often relies on heuristics. As a result, our method not only improves the training efficiency, but also mitigates the objective bias introduced by the sequential procedures that potentially lead to a suboptimal solution. Specifically, we formulate MTG as a fully differentiable pruning problem on an adaptive network architecture determined by an underlying Categorical distribution. To categorize N tasks into K groups (represented by K encoder branches), we initially set up KN task heads, where each branch connects to all N task heads to exploit the high-order task-affinity. Then, we gradually prune the KN heads down to N by learning a relaxed differentiable Categorical distribution, ensuring that each task is exclusively and uniquely categorized into only one branch. Extensive experiments on CelebA and Taskonomy datasets with detailed ablations show the promising performance and efficiency of our method. The codes are available at https://github.com/ethanygao/DMTG.
Authors: Xiao Siyao, Huang Libing, Zhang Shunsheng
Abstract: Multiplicative noise widely exists in radar images, medical images and other important fields' images. Compared to normal noises, multiplicative noise has a generally stronger effect on the visual expression of images. Aiming at the denoising problem of multiplicative noise, we linearize the nonlocal means algorithm with deep learning and propose a linear attention mechanism based deep nonlocal means filtering (LDNLM). Starting from the traditional nonlocal means filtering, we employ deep channel convolution neural networks to extract the information of the neighborhood matrix and obtain representation vectors of every pixel. Then we replace the similarity calculation and weighted averaging processes with the inner operations of the attention mechanism. To reduce the computational overhead, through the formula of similarity calculation and weighted averaging, we derive a nonlocal filter with linear complexity. Experiments on both simulated and real multiplicative noise demonstrate that the LDNLM is more competitive compared with the state-of-the-art methods. Additionally, we prove that the LDNLM possesses interpretability close to traditional NLM.
Authors: Suruchi Kumari, Aryan Das, Swalpa Kumar Roy, Indu Joshi, Pravendra Singh
Abstract: Traditional supervised 3D medical image segmentation models need voxel-level annotations, which require huge human effort, time, and cost. Semi-supervised learning (SSL) addresses this limitation of supervised learning by facilitating learning with a limited annotated and larger amount of unannotated training samples. However, state-of-the-art SSL models still struggle to fully exploit the potential of learning from unannotated samples. To facilitate effective learning from unannotated data, we introduce LLM-SegNet, which exploits a large language model (LLM) to integrate task-specific knowledge into our co-training framework. This knowledge aids the model in comprehensively understanding the features of the region of interest (ROI), ultimately leading to more efficient segmentation. Additionally, to further reduce erroneous segmentation, we propose a Unified Segmentation loss function. This loss function reduces erroneous segmentation by not only prioritizing regions where the model is confident in predicting between foreground or background pixels but also effectively addressing areas where the model lacks high confidence in predictions. Experiments on publicly available Left Atrium, Pancreas-CT, and Brats-19 datasets demonstrate the superior performance of LLM-SegNet compared to the state-of-the-art. Furthermore, we conducted several ablation studies to demonstrate the effectiveness of various modules and loss functions leveraged by LLM-SegNet.
Authors: Peng Xia, Kangyu Zhu, Haoran Li, Hongtu Zhu, Yun Li, Gang Li, Linjun Zhang, Huaxiu Yao
Abstract: The recent emergence of Medical Large Vision Language Models (Med-LVLMs) has enhanced medical diagnosis. However, current Med-LVLMs frequently encounter factual issues, often generating responses that do not align with established medical facts. Retrieval-Augmented Generation (RAG), which utilizes external knowledge, can improve the factual accuracy of these models but introduces two major challenges. First, limited retrieved contexts might not cover all necessary information, while excessive retrieval can introduce irrelevant and inaccurate references, interfering with the model's generation. Second, in cases where the model originally responds correctly, applying RAG can lead to an over-reliance on retrieved contexts, resulting in incorrect answers. To address these issues, we propose RULE, which consists of two components. First, we introduce a provably effective strategy for controlling factuality risk through the calibrated selection of the number of retrieved contexts. Second, based on samples where over-reliance on retrieved contexts led to errors, we curate a preference dataset to fine-tune the model, balancing its dependence on inherent knowledge and retrieved contexts for generation. We demonstrate the effectiveness of RULE on three medical VQA datasets, achieving an average improvement of 20.8% in factual accuracy. We publicly release our benchmark and code in https://github.com/richard-peng-xia/RULE.
Authors: Mohd Usama, Emma Nyman, Ulf Naslund, Christer Gronlund
Abstract: Deep learning has been used extensively for medical image analysis applications, assuming the training and test data adhere to the same probability distributions. However, a common challenge arises when dealing with medical images generated by different systems or even the same system with varying parameter settings. Such images often contain diverse textures and noise patterns, violating the assumption. Consequently, models trained on data from one machine or setting usually struggle to perform effectively on data from another. To address this issue in ultrasound images, we proposed a Generative Adversarial Network (GAN) based model in this paper. We formulated image harmonization and denoising tasks as an image-to-image translation task, wherein we modified the texture pattern and reduced noise in Carotid ultrasound images while keeping the image content (the anatomy) unchanged. The performance was evaluated using feature distribution and pixel-space similarity metrics. In addition, blood-to-tissue contrast and influence on computed risk markers (Gray scale median, GSM) were evaluated. The results showed that domain adaptation was achieved in both tasks (histogram correlation 0.920 and 0.844), as compared to no adaptation (0.890 and 0.707), and that the anatomy of the images was retained (structure similarity index measure of the arterial wall 0.71 and 0.80). In addition, the image noise level (contrast) did not change in the image harmonization task (-34.1 vs 35.2 dB) but was improved in the noise reduction task (-23.5 vs -46.7 dB). The model outperformed the CycleGAN in both tasks. Finally, the risk marker GSM increased by 7.6 (p<0.001) in task 1 but not in task 2. We conclude that domain translation models are powerful tools for ultrasound image improvement while retaining the underlying anatomy but that downstream calculations of risk markers may be affected.
Authors: Yuxuan Yan, Na Lu, Ruofan Yan
Abstract: Combining machine clustering with deep models has shown remarkable superiority in deep clustering. It modifies the data processing pipeline into two alternating phases: feature clustering and model training. However, such alternating schedule may lead to instability and computational burden issues. We propose a centerless clustering algorithm called Probability Aggregation Clustering (PAC) to proactively adapt deep learning technologies, enabling easy deployment in online deep clustering. PAC circumvents the cluster center and aligns the probability space and distribution space by formulating clustering as an optimization problem with a novel objective function. Based on the computation mechanism of the PAC, we propose a general online probability aggregation module to perform stable and flexible feature clustering over mini-batch data and further construct a deep visual clustering framework deep PAC (DPAC). Extensive experiments demonstrate that PAC has superior clustering robustness and performance and DPAC remarkably outperforms the state-of-the-art deep clustering methods.
Authors: Luzhe Huang, Xiongye Xiao, Shixuan Li, Jiawen Sun, Yi Huang, Aydogan Ozcan, Paul Bogdan
Abstract: The advance of diffusion-based generative models in recent years has revolutionized state-of-the-art (SOTA) techniques in a wide variety of image analysis and synthesis tasks, whereas their adaptation on image restoration, particularly within computational microscopy remains theoretically and empirically underexplored. In this research, we introduce a multi-scale generative model that enhances conditional image restoration through a novel exploitation of the Brownian Bridge process within wavelet domain. By initiating the Brownian Bridge diffusion process specifically at the lowest-frequency subband and applying generative adversarial networks at subsequent multi-scale high-frequency subbands in the wavelet domain, our method provides significant acceleration during training and sampling while sustaining a high image generation quality and diversity on par with SOTA diffusion models. Experimental results on various computational microscopy and imaging tasks confirm our method's robust performance and its considerable reduction in its sampling steps and time. This pioneering technique offers an efficient image restoration framework that harmonizes efficiency with quality, signifying a major stride in incorporating cutting-edge generative models into computational microscopy workflows.
Authors: Shaoyuan Chen, Linlin You, Rui Liu, Shuo Yu, Ahmed M. Abdelmoniem
Abstract: The training of large models, involving fine-tuning, faces the scarcity of high-quality data. Compared to the solutions based on centralized data centers, updating large models in the Internet of Things (IoT) faces challenges in coordinating knowledge from distributed clients by using their private and heterogeneous data. To tackle such a challenge, we propose KOALA (Federated Knowledge Transfer Fine-tuning Large Server Model with Resource-Constrained IoT Clients) to impel the training of large models in IoT. Since the resources obtained by IoT clients are limited and restricted, it is infeasible to locally execute large models and also update them in a privacy-preserving manner. Therefore, we leverage federated learning and knowledge distillation to update large models through collaboration with their small models, which can run locally at IoT clients to process their private data separately and enable large-small model knowledge transfer through iterative learning between the server and clients. Moreover, to support clients with similar or different computing capacities, KOALA is designed with two kinds of large-small model joint learning modes, namely to be homogeneous or heterogeneous. Experimental results demonstrate that compared to the conventional approach, our method can not only achieve similar training performance but also significantly reduce the need for local storage and computing power resources.
Authors: Chun-Mei Feng
Abstract: Aside from offering state-of-the-art performance in medical image generation, denoising diffusion probabilistic models (DPM) can also serve as a representation learner to capture semantic information and potentially be used as an image representation for downstream tasks, e.g., segmentation. However, these latent semantic representations rely heavily on labor-intensive pixel-level annotations as supervision, limiting the usability of DPM in medical image segmentation. To address this limitation, we propose an enhanced diffusion segmentation model, called TextDiff, that improves semantic representation through inexpensive medical text annotations, thereby explicitly establishing semantic representation and language correspondence for diffusion models. Concretely, TextDiff extracts intermediate activations of the Markov step of the reverse diffusion process in a pretrained diffusion model on large-scale natural images and learns additional expert knowledge by combining them with complementary and readily available diagnostic text information. TextDiff freezes the dual-branch multi-modal structure and mines the latent alignment of semantic features in diffusion models with diagnostic descriptions by only training the cross-attention mechanism and pixel classifier, making it possible to enhance semantic representation with inexpensive text. Extensive experiments on public QaTa-COVID19 and MoNuSeg datasets show that our TextDiff is significantly superior to the state-of-the-art multi-modal segmentation methods with only a few training samples.
Authors: Zirun Guo, Tao Jin, Zhou Zhao
Abstract: The development of multimodal models has significantly advanced multimodal sentiment analysis and emotion recognition. However, in real-world applications, the presence of various missing modality cases often leads to a degradation in the model's performance. In this work, we propose a novel multimodal Transformer framework using prompt learning to address the issue of missing modalities. Our method introduces three types of prompts: generative prompts, missing-signal prompts, and missing-type prompts. These prompts enable the generation of missing modality features and facilitate the learning of intra- and inter-modality information. Through prompt learning, we achieve a substantial reduction in the number of trainable parameters. Our proposed method outperforms other methods significantly across all evaluation metrics. Extensive experiments and ablation studies are conducted to demonstrate the effectiveness and robustness of our method, showcasing its ability to effectively handle missing modalities.
Authors: Stefan Horoi, Albert Manuel Orozco Camacho, Eugene Belilovsky, Guy Wolf
Abstract: Combining the predictions of multiple trained models through ensembling is generally a good way to improve accuracy by leveraging the different learned features of the models, however it comes with high computational and storage costs. Model fusion, the act of merging multiple models into one by combining their parameters reduces these costs but doesn't work as well in practice. Indeed, neural network loss landscapes are high-dimensional and non-convex and the minima found through learning are typically separated by high loss barriers. Numerous recent works have been focused on finding permutations matching one network features to the features of a second one, lowering the loss barrier on the linear path between them in parameter space. However, permutations are restrictive since they assume a one-to-one mapping between the different models' neurons exists. We propose a new model merging algorithm, CCA Merge, which is based on Canonical Correlation Analysis and aims to maximize the correlations between linear combinations of the model features. We show that our alignment method leads to better performances than past methods when averaging models trained on the same, or differing data splits. We also extend this analysis into the harder setting where more than 2 models are merged, and we find that CCA Merge works significantly better than past methods. Our code is publicly available at https://github.com/shoroi/align-n-merge
Authors: Chongjie Si, Xiaokang Yang, Wei Shen
Abstract: The rapid expansion of large foundation models within the pre-training and fine-tuning framework has underscored that larger models often yield better results. However, the scaling up of large foundation models has led to soaring costs in fine-tuning and parameter storage, rendering extensive adaptations impractical. This challenge has sparked the development of parameter-efficient fine-tuning (PEFT), which focuses on optimizing a select subset of parameters while keeping the rest fixed, significantly lowering computational and storage overheads. While recent years have witnessed a significant success in PEFT, a deep understanding of the fundamental principles behind these methods remains unexplored. To this end, here we take the first step to unify all approaches by dissecting them from a decomposition perspective. We initiate a comprehensive mathematical analysis of these methods, allowing us to delve deeply into their underlying mechanisms, and we explore the reasons behind the variations in performance among different techniques. Furthermore, inspired by our theoretical analysis, we introduce two novel PEFT methods alongside a simple yet effective framework designed to enhance the performance of PEFT techniques across various applications. Our empirical validations, conducted across multiple datasets, demonstrate the efficacy of these methods, showcasing both theoretical validity and practical performance improvements under the guidance of our analytical findings. We believe our work will deepen researchers' understanding of PEFT and other techniques, prompting further contemplation and advancing the research across the whole community.
Authors: Marina Dom\'inguez, Yordanka Velikova, Nassir Navab, Mohammad Farid Azampour
Abstract: Deep learning (DL) methods typically require large datasets to effectively learn data distributions. However, in the medical field, data is often limited in quantity, and acquiring labeled data can be costly. To mitigate this data scarcity, data augmentation techniques are commonly employed. Among these techniques, generative models play a pivotal role in expanding datasets. However, when it comes to ultrasound (US) imaging, the authenticity of generated data often diminishes due to the oversight of ultrasound physics. We propose a novel approach to improve the quality of generated US images by introducing a physics-based diffusion model that is specifically designed for this image modality. The proposed model incorporates an US-specific scheduler scheme that mimics the natural behavior of sound wave propagation in ultrasound imaging. Our analysis demonstrates how the proposed method aids in modeling the attenuation dynamics in US imaging. We present both qualitative and quantitative results based on standard generative model metrics, showing that our proposed method results in overall more plausible images. Our code is available at https://github.com/marinadominguez/diffusion-for-us-images
URLs: https://github.com/marinadominguez/diffusion-for-us-images
Authors: P. N. Karthikayan, Yoga Sri Varshan V, Hitesh Gupta Kattamuri, Umarani Jayaraman
Abstract: This paper presents dilated Residual Network (ResNet) models for disease classification from retinal fundus images. Dilated convolution filters are used to replace normal convolution filters in the higher layers of the ResNet model (dilated ResNet) in order to improve the receptive field compared to the normal ResNet model for disease classification. This study introduces computer-assisted diagnostic tools that employ deep learning, enhanced with explainable AI techniques. These techniques aim to make the tool's decision-making process transparent, thereby enabling medical professionals to understand and trust the AI's diagnostic decision. They are particularly relevant in today's healthcare landscape, where there is a growing demand for transparency in AI applications to ensure their reliability and ethical use. The dilated ResNet is used as a replacement for the normal ResNet to enhance the classification accuracy of retinal eye diseases and reduce the required computing time. The dataset used in this work is the Ocular Disease Intelligent Recognition (ODIR) dataset which is a structured ophthalmic database with eight classes covering most of the common retinal eye diseases. The evaluation metrics used in this work include precision, recall, accuracy, and F1 score. In this work, a comparative study has been made between normal ResNet models and dilated ResNet models on five variants namely ResNet-18, ResNet-34, ResNet-50, ResNet-101, and ResNet-152. The dilated ResNet model shows promising results as compared to normal ResNet with an average F1 score of 0.71, 0.70, 0.69, 0.67, and 0.70 respectively for the above respective variants in ODIR multiclass disease classification.
Authors: Fangqiang Xu, Wenxuan Tu, Fan Feng, Malitha Gunawardhana, Jiayuan Yang, Yun Gu, Jichao Zhao
Abstract: Left atrial (LA) segmentation is a crucial technique for irregular heartbeat (i.e., atrial fibrillation) diagnosis. Most current methods for LA segmentation strictly assume that the input data is acquired using object-oriented center cropping, while this assumption may not always hold in practice due to the high cost of manual object annotation. Random cropping is a straightforward data pre-processing approach. However, it 1) introduces significant irregularities and incompleteness in the input data and 2) disrupts the coherence and continuity of object boundary regions. To tackle these issues, we propose a novel Dynamic Position transformation and Boundary refinement Network (DPBNet). The core idea is to dynamically adjust the relative position of irregular targets to construct their contextual relationships and prioritize difficult boundary pixels to enhance foreground-background distinction. Specifically, we design a shuffle-then-reorder attention module to adjust the position of disrupted objects in the latent space using dynamic generation ratios, such that the vital dependencies among these random cropping targets could be well captured and preserved. Moreover, to improve the accuracy of boundary localization, we introduce a dual fine-grained boundary loss with scenario-adaptive weights to handle the ambiguity of the dual boundary at a fine-grained level, promoting the clarity and continuity of the obtained results. Extensive experimental results on benchmark dataset have demonstrated that DPBNet consistently outperforms existing state-of-the-art methods.
Authors: Boyang Wang, Nikhil Sridhar, Chao Feng, Mark Van der Merwe, Adam Fishman, Nima Fazeli, Jeong Joon Park
Abstract: We propose a robot learning method for communicating, planning, and executing a wide range of tasks, dubbed This&That. We achieve robot planning for general tasks by leveraging the power of video generative models trained on internet-scale data containing rich physical and semantic context. In this work, we tackle three fundamental challenges in video-based planning: 1) unambiguous task communication with simple human instructions, 2) controllable video generation that respects user intents, and 3) translating visual planning into robot actions. We propose language-gesture conditioning to generate videos, which is both simpler and clearer than existing language-only methods, especially in complex and uncertain environments. We then suggest a behavioral cloning design that seamlessly incorporates the video plans. This&That demonstrates state-of-the-art effectiveness in addressing the above three challenges, and justifies the use of video generation as an intermediate representation for generalizable task planning and execution. Project website: https://cfeng16.github.io/this-and-that/.
Authors: Inye Na, Jonghun Kim, Eun Sook Ko, Hyunjin Park
Abstract: Motivated by the question, "Can we generate tumors with desired attributes?'' this study leverages radiomics features to explore the feasibility of generating synthetic tumor images. Characterized by its low-dimensional yet biologically meaningful markers, radiomics bridges the gap between complex medical imaging data and actionable clinical insights. We present RadiomicsFill-Mammo, the first of the RadiomicsFill series, an innovative technique that generates realistic mammogram mass images mirroring specific radiomics attributes using masked images and opposite breast images, leveraging a recent stable diffusion model. This approach also allows for the incorporation of essential clinical variables, such as BI-RADS and breast density, alongside radiomics features as conditions for mass generation. Results indicate that RadiomicsFill-Mammo effectively generates diverse and realistic tumor images based on various radiomics conditions. Results also demonstrate a significant improvement in mass detection capabilities, leveraging RadiomicsFill-Mammo as a strategy to generate simulated samples. Furthermore, RadiomicsFill-Mammo not only advances medical imaging research but also opens new avenues for enhancing treatment planning and tumor simulation. Our code is available at https://github.com/nainye/RadiomicsFill.
Authors: Martin B\"uchner, Simon Dorer, Abhinav Valada
Abstract: The robust and safe operation of automated vehicles underscores the critical need for detailed and accurate topological maps. At the heart of this requirement is the construction of lane graphs, which provide essential information on lane connectivity, vital for navigating complex urban environments autonomously. While transformer-based models have been effective in creating map topologies from vehicle-mounted sensor data, their potential for generating such graphs from aerial imagery remains untapped. This work introduces a novel approach to generating successor lane graphs from aerial imagery, utilizing the advanced capabilities of transformer models. We frame successor lane graphs as a collection of maximal length paths and predict them using a Detection Transformer (DETR) architecture. We demonstrate the efficacy of our method through extensive experiments on the diverse and large-scale UrbanLaneGraph dataset, illustrating its accuracy in generating successor lane graphs and highlighting its potential for enhancing autonomous vehicle navigation in complex environments.
Authors: Chunwei Tian, Menghua Zheng, Chia-Wen Lin, Zhiwu Li, David Zhang
Abstract: Deep networks can usually depend on extracting more structural information to improve denoising results. However, they may ignore correlation between pixels from an image to pursue better denoising performance. Window transformer can use long- and short-distance modeling to interact pixels to address mentioned problem. To make a tradeoff between distance modeling and denoising time, we propose a heterogeneous window transformer (HWformer) for image denoising. HWformer first designs heterogeneous global windows to capture global context information for improving denoising effects. To build a bridge between long and short-distance modeling, global windows are horizontally and vertically shifted to facilitate diversified information without increasing denoising time. To prevent the information loss phenomenon of independent patches, sparse idea is guided a feed-forward network to extract local information of neighboring patches. The proposed HWformer only takes 30% of popular Restormer in terms of denoising time.
Authors: Qi Zhang, Zhihao Lin, Arnoud Visser
Abstract: The ICRA conference is celebrating its $40^{th}$ anniversary in Rotterdam in September 2024, with as highlight the Happy Birthday ICRA Party at the iconic Holland America Line Cruise Terminal. One month later the IROS conference will take place, which will include the Earth Rover Challenge. In this challenge open-world autonomous navigation models are studied truly open-world settings. As part of the Earth Rover Challenge several real-world navigation sets in several cities world-wide, like Auckland, Australia and Wuhan, China. The only dataset recorded in the Netherlands is the small village Oudewater. The proposal is to record a dataset with the robot used in the Earth Rover Challenge in Rotterdam, in front of the Holland America Line Cruise Terminal, before the festivities of the Happy Birthday ICRA Party start.
Authors: Kun Wu, Zixu Wang, Xiulong Yang, Yangyang Chen, Zhenqi Han, Jialu Zhang, Lizhuang Liu
Abstract: As the primary mRNA delivery vehicles, ionizable lipid nanoparticles (LNPs) exhibit excellent safety, high transfection efficiency, and strong immune response induction. However, the screening process for LNPs is time-consuming and costly. To expedite the identification of high-transfection-efficiency mRNA drug delivery systems, we propose an explainable LNPs transfection efficiency prediction model, called TransMA. TransMA employs a multi-modal molecular structure fusion architecture, wherein the fine-grained atomic spatial relationship extractor named molecule 3D Transformer captures three-dimensional spatial features of the molecule, and the coarse-grained atomic sequence extractor named molecule Mamba captures one-dimensional molecular features. We design the mol-attention mechanism block, enabling it to align coarse and fine-grained atomic features and captures relationships between atomic spatial and sequential structures. TransMA achieves state-of-the-art performance in predicting transfection efficiency using the scaffold and cliff data splitting methods on the current largest LNPs dataset, including Hela and RAW cell lines. Moreover, we find that TransMA captures the relationship between subtle structural changes and significant transfection efficiency variations, providing valuable insights for LNPs design. Additionally, TransMA's predictions on external transfection efficiency data maintain a consistent order with actual transfection efficiencies, demonstrating its robust generalization capability. The code, model and data are made publicly available at https://github.com/wklix/TransMA/tree/master. We hope that high-accuracy transfection prediction models in the future can aid in LNPs design and initial screening, thereby assisting in accelerating the mRNA design process.
Authors: Yutong Zhang, Yi Pan, Tianyang Zhong, Peixin Dong, Kangni Xie, Yuxiao Liu, Hanqi Jiang, Zhengliang Liu, Shijie Zhao, Tuo Zhang, Xi Jiang, Dinggang Shen, Tianming Liu, Xin Zhang
Abstract: Medical images and radiology reports are crucial for diagnosing medical conditions, highlighting the importance of quantitative analysis for clinical decision-making. However, the diversity and cross-source heterogeneity of these data challenge the generalizability of current data-mining methods. Multimodal large language models (MLLMs) have recently transformed many domains, significantly affecting the medical field. Notably, Gemini-Vision-series (Gemini) and GPT-4-series (GPT-4) models have epitomized a paradigm shift in Artificial General Intelligence (AGI) for computer vision, showcasing their potential in the biomedical domain. In this study, we evaluated the performance of the Gemini, GPT-4, and 4 popular large models for an exhaustive evaluation across 14 medical imaging datasets, including 5 medical imaging categories (dermatology, radiology, dentistry, ophthalmology, and endoscopy), and 3 radiology report datasets. The investigated tasks encompass disease classification, lesion segmentation, anatomical localization, disease diagnosis, report generation, and lesion detection. Our experimental results demonstrated that Gemini-series models excelled in report generation and lesion detection but faces challenges in disease classification and anatomical localization. Conversely, GPT-series models exhibited proficiency in lesion segmentation and anatomical localization but encountered difficulties in disease diagnosis and lesion detection. Additionally, both the Gemini series and GPT series contain models that have demonstrated commendable generation efficiency. While both models hold promise in reducing physician workload, alleviating pressure on limited healthcare resources, and fostering collaboration between clinical practitioners and artificial intelligence technologies, substantial enhancements and comprehensive validations remain imperative before clinical deployment.
Authors: Nataliia Molchanova, Alessandro Cagol, Pedro M. Gordaliza, Mario Ocampo-Pineda, Po-Jui Lu, Matthias Weigel, Xinjie Chen, Adrien Depeursinge, Cristina Granziera, Henning M\"uller, Meritxell Bach Cuadra
Abstract: Uncertainty quantification (UQ) has become critical for evaluating the reliability of artificial intelligence systems, especially in medical image segmentation. This study addresses the interpretability of instance-wise uncertainty values in deep learning models for focal lesion segmentation in magnetic resonance imaging, specifically cortical lesion (CL) segmentation in multiple sclerosis. CL segmentation presents several challenges, including the complexity of manual segmentation, high variability in annotation, data scarcity, and class imbalance, all of which contribute to aleatoric and epistemic uncertainty. We explore how UQ can be used not only to assess prediction reliability but also to provide insights into model behavior, detect biases, and verify the accuracy of UQ methods. Our research demonstrates the potential of instance-wise uncertainty values to offer post hoc global model explanations, serving as a sanity check for the model. The implementation is available at https://github.com/NataliiaMolch/interpret-lesion-unc.
URLs: https://github.com/NataliiaMolch/interpret-lesion-unc.
Authors: Qi Li, Ziyi Shen, Qianye Yang, Dean C. Barratt, Matthew J. Clarkson, Tom Vercauteren, Yipeng Hu
Abstract: Reconstructing 2D freehand Ultrasound (US) frames into 3D space without using a tracker has recently seen advances with deep learning. Predicting good frame-to-frame rigid transformations is often accepted as the learning objective, especially when the ground-truth labels from spatial tracking devices are inherently rigid transformations. Motivated by a) the observed nonrigid deformation due to soft tissue motion during scanning, and b) the highly sensitive prediction of rigid transformation, this study investigates the methods and their benefits in predicting nonrigid transformations for reconstructing 3D US. We propose a novel co-optimisation algorithm for simultaneously estimating rigid transformations among US frames, supervised by ground-truth from a tracker, and a nonrigid deformation, optimised by a regularised registration network. We show that these two objectives can be either optimised using meta-learning or combined by weighting. A fast scattered data interpolation is also developed for enabling frequent reconstruction and registration of non-parallel US frames, during training. With a new data set containing over 357,000 frames in 720 scans, acquired from 60 subjects, the experiments demonstrate that, due to an expanded thus easier-to-optimise solution space, the generalisation is improved with the added deformation estimation, with respect to the rigid ground-truth. The global pixel reconstruction error (assessing accumulative prediction) is lowered from 18.48 to 16.51 mm, compared with baseline rigid-transformation-predicting methods. Using manually identified landmarks, the proposed co-optimisation also shows potentials in compensating nonrigid tissue motion at inference, which is not measurable by tracker-provided ground-truth. The code and data used in this paper are made publicly available at https://github.com/QiLi111/NR-Rec-FUS.
Authors: Ioannis Tsiamas, Santiago Pascual, Chunghsin Yeh, Joan Serr\`a
Abstract: Contrastive learning has emerged as a powerful technique in audio-visual representation learning, leveraging the natural co-occurrence of audio and visual modalities in extensive web-scale video datasets to achieve significant advancements. However, conventional contrastive audio-visual learning methodologies often rely on aggregated representations derived through temporal aggregation, which neglects the intrinsic sequential nature of the data. This oversight raises concerns regarding the ability of standard approaches to capture and utilize fine-grained information within sequences, information that is vital for distinguishing between semantically similar yet distinct examples. In response to this limitation, we propose sequential contrastive audio-visual learning (SCAV), which contrasts examples based on their non-aggregated representation space using sequential distances. Retrieval experiments with the VGGSound and Music datasets demonstrate the effectiveness of SCAV, showing 2-3x relative improvements against traditional aggregation-based contrastive learning and other methods from the literature. We also show that models trained with SCAV exhibit a high degree of flexibility regarding the metric employed for retrieval, allowing them to operate on a spectrum of efficiency-accuracy trade-offs, potentially making them applicable in multiple scenarios, from small- to large-scale retrieval.
Authors: Yinsong Xu, Yipei Wang, Ziyi Shen, Iani J. M. B. Gayo, Natasha Thorley, Shonit Punwani, Aidong Men, Dean Barratt, Qingchao Chen, Yipeng Hu
Abstract: The Gleason groups serve as the primary histological grading system for prostate cancer, providing crucial insights into the cancer's potential for growth and metastasis. In clinical practice, pathologists determine the Gleason groups based on specimens obtained from ultrasound-guided biopsies. In this study, we investigate the feasibility of directly estimating the Gleason groups from MRI scans to reduce otherwise required biopsies. We identify two characteristics of this task, ordinality and the resulting dependent yet unknown variances between Gleason groups. In addition to the inter- / intra- observer variability in a multi-step Gleason scoring process based on the interpretation of Gleason patterns, our MR-based prediction is also subject to specimen sampling variance and, to a lesser degree, varying MR imaging protocols. To address this challenge, we propose a novel Poisson ordinal network (PON). PONs model the prediction using a Poisson distribution and leverages Poisson encoding and Poisson focal loss to capture a learnable dependency between ordinal classes (here, Gleason groups), rather than relying solely on the numerical ground-truth (e.g. Gleason Groups 1-5 or Gleason Scores 6-10). To improve this modelling efficacy, PONs also employ contrastive learning with a memory bank to regularise intra-class variance, decoupling the memory requirement of contrast learning from the batch size. Experimental results based on the images labelled by saturation biopsies from 265 prior-biopsy-blind patients, across two tasks demonstrate the superiority and effectiveness of our proposed method.
Authors: Pranab Sahoo, Ashutosh Tripathi, Sriparna Saha, Samrat Mondal
Abstract: Despite recent advancements in federated learning (FL) for medical image diagnosis, addressing data heterogeneity among clients remains a significant challenge for practical implementation. A primary hurdle in FL arises from the non-IID nature of data samples across clients, which typically results in a decline in the performance of the aggregated global model. In this study, we introduce FedMRL, a novel federated multi-agent deep reinforcement learning framework designed to address data heterogeneity. FedMRL incorporates a novel loss function to facilitate fairness among clients, preventing bias in the final global model. Additionally, it employs a multi-agent reinforcement learning (MARL) approach to calculate the proximal term $(\mu)$ for the personalized local objective function, ensuring convergence to the global optimum. Furthermore, FedMRL integrates an adaptive weight adjustment method using a Self-organizing map (SOM) on the server side to counteract distribution shifts among clients' local data distributions. We assess our approach using two publicly available real-world medical datasets, and the results demonstrate that FedMRL significantly outperforms state-of-the-art techniques, showing its efficacy in addressing data heterogeneity in federated learning. The code can be found here~{\url{https://github.com/Pranabiitp/FedMRL}}.
Authors: Ignacio Garrido Botella, Ignacio Arranz \'Agueda, Juan Carlos Armenteros Carmona, Oleg Vorontsov, Fernando Bay\'on Robledo, Adri\'an Alonso Barriuso
Abstract: Accurate identification, localization, and segregation of teeth from Cone Beam Computed Tomography (CBCT) images are essential for analyzing dental pathologies. Modeling an individual tooth can be challenging and intricate to accomplish, especially when fillings and other restorations introduce artifacts. This paper proposes a method for automatically detecting, identifying, and extracting teeth from CBCT images. Our approach involves dividing the three-dimensional images into axial slices for image detection. Teeth are pinpointed and labeled using a single-stage object detector. Subsequently, bounding boxes are delineated and identified to create three-dimensional representations of each tooth. The proposed solution has been successfully integrated into the dental analysis tool Dentomo.
Authors: Nick John Eliopoulos, Purvish Jajal, James Davis, Gaowen Liu, George K. Thiravathukal, Yung-Hsiang Lu
Abstract: This paper investigates how to efficiently deploy transformer-based neural networks on edge devices. Recent methods reduce the latency of transformer neural networks by removing or merging tokens, with small accuracy degradation. However, these methods are not designed with edge device deployment in mind, and do not leverage information about the hardware characteristics to improve efficiency. First, we show that the relationship between latency and workload size is governed by the GPU tail-effect. This relationship is used to create a token pruning schedule tailored for a pre-trained model and device pair. Second, we demonstrate a training-free token pruning method utilizing this relationship. This method achieves accuracy-latency trade-offs in a hardware aware manner. We show that for single batch inference, other methods may actually increase latency by 18.6-30.3% with respect to baseline, while we can reduce it by 9%. For similar latency (within 5.2%) across devices we achieve 78.6%-84.5% ImageNet1K accuracy, while the state-of-the-art, Token Merging, achieves 45.8%-85.4%.
Authors: Bidur Khanal, Tianhong Dai, Binod Bhattarai, Cristian Linte
Abstract: The robustness of supervised deep learning-based medical image classification is significantly undermined by label noise. Although several methods have been proposed to enhance classification performance in the presence of noisy labels, they face some challenges: 1) a struggle with class-imbalanced datasets, leading to the frequent overlooking of minority classes as noisy samples; 2) a singular focus on maximizing performance using noisy datasets, without incorporating experts-in-the-loop for actively cleaning the noisy labels. To mitigate these challenges, we propose a two-phase approach that combines Learning with Noisy Labels (LNL) and active learning. This approach not only improves the robustness of medical image classification in the presence of noisy labels, but also iteratively improves the quality of the dataset by relabeling the important incorrect labels, under a limited annotation budget. Furthermore, we introduce a novel Variance of Gradients approach in LNL phase, which complements the loss-based sample selection by also sampling under-represented samples. Using two imbalanced noisy medical classification datasets, we demonstrate that that our proposed technique is superior to its predecessors at handling class imbalance by not misidentifying clean samples from minority classes as mostly noisy samples.
Authors: Luigi Capogrosso, Enrico Fraccaroli, Samarjit Chakraborty, Franco Fummi, Marco Cristani
Abstract: Split Computing (SC), where a Deep Neural Network (DNN) is intelligently split with a part of it deployed on an edge device and the rest on a remote server is emerging as a promising approach. It allows the power of DNNs to be leveraged for latency-sensitive applications that do not allow the entire DNN to be deployed remotely, while not having sufficient computation bandwidth available locally. In many such embedded systems scenarios, such as those in the automotive domain, computational resource constraints also necessitate Multi-Task Learning (MTL), where the same DNN is used for multiple inference tasks instead of having dedicated DNNs for each task, which would need more computing bandwidth. However, how to partition such a multi-tasking DNN to be deployed within a SC framework has not been sufficiently studied. This paper studies this problem, and MTL-Split, our novel proposed architecture, shows encouraging results on both synthetic and real-world data. The source code is available at https://github.com/intelligolabs/MTL-Split.
Authors: Shuang Xu, Qiao Ke, Jiangjun Peng, Xiangyong Cao, Zixiang Zhao
Abstract: This paper introduces a novel paradigm for hyperspectral image (HSI) denoising, which is termed \textit{pan-denoising}. In a given scene, panchromatic (PAN) images capture similar structures and textures to HSIs but with less noise. This enables the utilization of PAN images to guide the HSI denoising process. Consequently, pan-denoising, which incorporates an additional prior, has the potential to uncover underlying structures and details beyond the internal information modeling of traditional HSI denoising methods. However, the proper modeling of this additional prior poses a significant challenge. To alleviate this issue, the paper proposes a novel regularization term, Panchromatic Weighted Representation Coefficient Total Variation (PWRCTV). It employs the gradient maps of PAN images to automatically assign different weights of TV regularization for each pixel, resulting in larger weights for smooth areas and smaller weights for edges. This regularization forms the basis of a pan-denoising model, which is solved using the Alternating Direction Method of Multipliers. Extensive experiments on synthetic and real-world datasets demonstrate that PWRCTV outperforms several state-of-the-art methods in terms of metrics and visual quality. Furthermore, an HSI classification experiment confirms that PWRCTV, as a preprocessing method, can enhance the performance of downstream classification tasks. The code and data are available at https://github.com/shuangxu96/PWRCTV.
Authors: Siva Krishna Ravipati, Ehsan Latif, Ramviyas Parasuraman, Suchendra M. Bhandarkar
Abstract: Classification of different object surface material types can play a significant role in the decision-making algorithms for mobile robots and autonomous vehicles. RGB-based scene-level semantic segmentation has been well-addressed in the literature. However, improving material recognition using the depth modality and its integration with SLAM algorithms for 3D semantic mapping could unlock new potential benefits in the robotics perception pipeline. To this end, we propose a complementarity-aware deep learning approach for RGB-D-based material classification built on top of an object-oriented pipeline. The approach further integrates the ORB-SLAM2 method for 3D scene mapping with multiscale clustering of the detected material semantics in the point cloud map generated by the visual SLAM algorithm. Extensive experimental results with existing public datasets and newly contributed real-world robot datasets demonstrate a significant improvement in material classification and 3D clustering accuracy compared to state-of-the-art approaches for 3D semantic scene mapping.
Authors: Jingmao Cui, Donglai Tao, Linmi Tao, Ruiyang Liu, Yu Cheng
Abstract: The prevailing approach to embedding prior knowledge within convolutional layers typically includes the design of steerable kernels or their modulation using designated kernel banks. In this study, we introduce the Analytic Convolutional Layer (ACL), an innovative model-driven convolutional layer, which is a mosaic of analytical convolution kernels (ACKs) and traditional convolution kernels. ACKs are characterized by mathematical functions governed by analytic kernel parameters (AKPs) learned in training process. Learnable AKPs permit the adaptive update of incorporated knowledge to align with the features representation of data. Our extensive experiments demonstrate that the ACLs not only have a remarkable capacity for feature representation with a reduced number of parameters but also attain increased reliability through the analytical formulation of ACKs. Furthermore, ACLs offer a means for neural network interpretation, thereby paving the way for the intrinsic interpretability of neural network. The source code will be published in company with the paper.
Authors: Peizhuo Li, Tuanfeng Y. Wang, Timur Levent Kesdogan, Duygu Ceylan, Olga Sorkine-Hornung
Abstract: Data driven and learning based solutions for modeling dynamic garments have significantly advanced, especially in the context of digital humans. However, existing approaches often focus on modeling garments with respect to a fixed parametric human body model and are limited to garment geometries that were seen during training. In this work, we take a different approach and model the dynamics of a garment by exploiting its local interactions with the underlying human body. Specifically, as the body moves, we detect local garment-body collisions, which drive the deformation of the garment. At the core of our approach is a mesh-agnostic garment representation and a manifold-aware transformer network design, which together enable our method to generalize to unseen garment and body geometries. We evaluate our approach on a wide variety of garment types and motion sequences and provide competitive qualitative and quantitative results with respect to the state of the art.
Authors: Lucas W. Remedios, Shunxing Bao, Samuel W. Remedios, Ho Hin Lee, Leon Y. Cai, Thomas Li, Ruining Deng, Nancy R. Newlin, Adam M. Saunders, Can Cui, Jia Li, Qi Liu, Ken S. Lau, Joseph T. Roland, Mary K Washington, Lori A. Coburn, Keith T. Wilson, Yuankai Huo, Bennett A. Landman
Abstract: Understanding the way cells communicate, co-locate, and interrelate is essential to furthering our understanding of how the body functions. H&E is widely available, however, cell subtyping often requires expert knowledge and the use of specialized stains. To reduce the annotation burden, AI has been proposed for the classification of cells on H&E. For example, the recent Colon Nucleus Identification and Classification (CoNIC) Challenge focused on labeling 6 cell types on H&E of the colon. However, the CoNIC Challenge was unable to classify epithelial subtypes (progenitor, enteroendocrine, goblet), lymphocyte subtypes (B, helper T, cytotoxic T), and connective subtypes (fibroblasts). We use inter-modality learning to label previously un-labelable cell types on H&E. We take advantage of multiplexed immunofluorescence (MxIF) histology to label 14 cell subclasses. We performed style transfer on the same MxIF tissues to synthesize realistic virtual H&E which we paired with the MxIF-derived cell subclassification labels. We evaluated the efficacy of using a supervised learning scheme where the input was realistic-quality virtual H&E and the labels were MxIF-derived cell subclasses. We assessed our model on private virtual H&E and public real H&E. On virtual H&E, we were able to classify helper T cells and epithelial progenitors with positive predictive values of $0.34 \pm 0.15$ (prevalence $0.03 \pm 0.01$) and $0.47 \pm 0.1$ (prevalence $0.07 \pm 0.02$) respectively, when using ground truth centroid information. On real H&E we could classify helper T cells and epithelial progenitors with upper bound positive predictive values of $0.43 \pm 0.03$ (parent class prevalence 0.21) and $0.94 \pm 0.02$ (parent class prevalence 0.49) when using ground truth centroid information. This is the first work to provide cell type classification for helper T and epithelial progenitor nuclei on H&E.
Authors: Jorge da Silva Goncalves, Laura Manduchi, Moritz Vandenhirtz, Julia Vogt
Abstract: This paper introduces Diffuse-TreeVAE, a deep generative model that integrates hierarchical clustering into the framework of Denoising Diffusion Probabilistic Models (DDPMs). The proposed approach generates new images by sampling from a root embedding of a learned latent tree VAE-based structure, it then propagates through hierarchical paths, and utilizes a second-stage DDPM to refine and generate distinct, high-quality images for each data cluster. The result is a model that not only improves image clarity but also ensures that the generated samples are representative of their respective clusters, addressing the limitations of previous VAE-based methods and advancing the state of clustering-based generative modeling.
Authors: Ethan Chern, Jiadi Su, Yan Ma, Pengfei Liu
Abstract: Previous open-source large multimodal models (LMMs) have faced several limitations: (1) they often lack native integration, requiring adapters to align visual representations with pre-trained large language models (LLMs); (2) many are restricted to single-modal generation; (3) while some support multimodal generation, they rely on separate diffusion models for visual modeling and generation. To mitigate these limitations, we present Anole, an open, autoregressive, native large multimodal model for interleaved image-text generation. We build Anole from Meta AI's Chameleon, adopting an innovative fine-tuning strategy that is both data-efficient and parameter-efficient. Anole demonstrates high-quality, coherent multimodal generation capabilities. We have open-sourced our model, training framework, and instruction tuning data.
Authors: Yan Xia, Ran Ding, Ziyuan Qin, Guanqi Zhan, Kaichen Zhou, Long Yang, Hao Dong, Daniel Cremers
Abstract: Recent advances in predicting 6D grasp poses from a single depth image have led to promising performance in robotic grasping. However, previous grasping models face challenges in cluttered environments where nearby objects impact the target object's grasp. In this paper, we first establish a new benchmark dataset for TARget-driven Grasping under Occlusions, named TARGO. We make the following contributions: 1) We are the first to study the occlusion level of grasping. 2) We set up an evaluation benchmark consisting of large-scale synthetic data and part of real-world data, and we evaluated five grasp models and found that even the current SOTA model suffers when the occlusion level increases, leaving grasping under occlusion still a challenge. 3) We also generate a large-scale training dataset via a scalable pipeline, which can be used to boost the performance of grasping under occlusion and generalized to the real world. 4) We further propose a transformer-based grasping model involving a shape completion module, termed TARGO-Net, which performs most robustly as occlusion increases. Our benchmark dataset can be found at https://TARGO-benchmark.github.io/.
Authors: Yunhao Luo, Chen Sun, Joshua B. Tenenbaum, Yilun Du
Abstract: Effective motion planning in high dimensional spaces is a long-standing open problem in robotics. One class of traditional motion planning algorithms corresponds to potential-based motion planning. An advantage of potential based motion planning is composability -- different motion constraints can be easily combined by adding corresponding potentials. However, constructing motion paths from potentials requires solving a global optimization across configuration space potential landscape, which is often prone to local minima. We propose a new approach towards learning potential based motion planning, where we train a neural network to capture and learn an easily optimizable potentials over motion planning trajectories. We illustrate the effectiveness of such approach, significantly outperforming both classical and recent learned motion planning approaches and avoiding issues with local minima. We further illustrate its inherent composability, enabling us to generalize to a multitude of different motion constraints.
Authors: R. Gnana Praveen, Eric Granger, Patrick Cardinal
Abstract: Automatic pain assessment has an important potential diagnostic value for populations that are incapable of articulating their pain experiences. As one of the dominating nonverbal channels for eliciting pain expression events, facial expressions has been widely investigated for estimating the pain intensity of individual. However, using state-of-the-art deep learning (DL) models in real-world pain estimation applications poses several challenges related to the subjective variations of facial expressions, operational capture conditions, and lack of representative training videos with labels. Given the cost of annotating intensity levels for every video frame, we propose a weakly-supervised domain adaptation (WSDA) technique that allows for training 3D CNNs for spatio-temporal pain intensity estimation using weakly labeled videos, where labels are provided on a periodic basis. In particular, WSDA integrates multiple instance learning into an adversarial deep domain adaptation framework to train an Inflated 3D-CNN (I3D) model such that it can accurately estimate pain intensities in the target operational domain. The training process relies on weak target loss, along with domain loss and source loss for domain adaptation of the I3D model. Experimental results obtained using labeled source domain RECOLA videos and weakly-labeled target domain UNBC-McMaster videos indicate that the proposed deep WSDA approach can achieve significantly higher level of sequence (bag)-level and frame (instance)-level pain localization accuracy than related state-of-the-art approaches.
Authors: R. Gnana Praveen, Eric Granger, Patrick Cardinal
Abstract: Estimation of pain intensity from facial expressions captured in videos has an immense potential for health care applications. Given the challenges related to subjective variations of facial expressions, and operational capture conditions, the accuracy of state-of-the-art DL models for recognizing facial expressions may decline. Domain adaptation has been widely explored to alleviate the problem of domain shifts that typically occur between video data captured across various source and target domains. Moreover, given the laborious task of collecting and annotating videos, and subjective bias due to ambiguity among adjacent intensity levels, weakly-supervised learning is gaining attention in such applications. State-of-the-art WSL models are typically formulated as regression problems, and do not leverage the ordinal relationship among pain intensity levels, nor temporal coherence of multiple consecutive frames. This paper introduces a new DL model for weakly-supervised DA with ordinal regression that can be adapted using target domain videos with coarse labels provided on a periodic basis. The WSDA-OR model enforces ordinal relationships among intensity levels assigned to target sequences, and associates multiple relevant frames to sequence-level labels. In particular, it learns discriminant and domain-invariant feature representations by integrating multiple instance learning with deep adversarial DA, where soft Gaussian labels are used to efficiently represent the weak ordinal sequence-level labels from target domain. The proposed approach was validated using RECOLA video dataset as fully-labeled source domain data, and UNBC-McMaster shoulder pain video dataset as weakly-labeled target domain data. We have also validated WSDA-OR on BIOVID and Fatigue datasets for sequence level estimation.
Authors: R. Gnana Praveen, Eric Granger, Patrick Cardinal
Abstract: In the recent years, there has been a shift in facial behavior analysis from the laboratory-controlled conditions to the challenging in-the-wild conditions due to the superior performance of deep learning based approaches for many real world applications.However, the performance of deep learning approaches relies on the amount of training data. One of the major problems with data acquisition is the requirement of annotations for large amount of training data. Labeling process of huge training data demands lot of human support with strong domain expertise for facial expressions or action units, which is difficult to obtain in real-time environments.Moreover, labeling process is highly vulnerable to ambiguity of expressions or action units, especially for intensities due to the bias induced by the domain experts. Therefore, there is an imperative need to address the problem of facial behavior analysis with weak annotations. In this paper, we provide a comprehensive review of weakly supervised learning (WSL) approaches for facial behavior analysis with both categorical as well as dimensional labels along with the challenges and potential research directions associated with it. First, we introduce various types of weak annotations in the context of facial behavior analysis and the corresponding challenges associated with it. We then systematically review the existing state-of-the-art approaches and provide a taxonomy of these approaches along with their insights and limitations. In addition, widely used data-sets in the reviewed literature and the performance of these approaches along with evaluation principles are summarized. Finally, we discuss the remaining challenges and opportunities along with the potential research directions in order to apply facial behavior analysis with weak labels in real life situations.
Authors: Faegheh Sardari, Bj\"orn Ommer, Majid Mirmehdi
Abstract: Most recent view-invariant action recognition and performance assessment approaches rely on a large amount of annotated 3D skeleton data to extract view-invariant features. However, acquiring 3D skeleton data can be cumbersome, if not impractical, in in-the-wild scenarios. To overcome this problem, we present a novel unsupervised approach that learns to extract view-invariant 3D human pose representation from a 2D image without using 3D joint data. Our model is trained by exploiting the intrinsic view-invariant properties of human pose between simultaneous frames from different viewpoints and their equivariant properties between augmented frames from the same viewpoint. We evaluate the learned view-invariant pose representations for two downstream tasks. We perform comparative experiments that show improvements on the state-of-the-art unsupervised cross-view action classification accuracy on NTU RGB+D by a significant margin, on both RGB and depth images. We also show the efficiency of transferring the learned representations from NTU RGB+D to obtain the first ever unsupervised cross-view and cross-subject rank correlation results on the multi-view human movement quality dataset, QMAR, and marginally improve on the-state-of-the-art supervised results for this dataset. We also carry out ablation studies to examine the contributions of the different components of our proposed network.
Authors: R. Gnana Praveen, Eric Granger, Patrick Cardinal
Abstract: Multimodal analysis has recently drawn much interest in affective computing, since it can improve the overall accuracy of emotion recognition over isolated uni-modal approaches. The most effective techniques for multimodal emotion recognition efficiently leverage diverse and complimentary sources of information, such as facial, vocal, and physiological modalities, to provide comprehensive feature representations. In this paper, we focus on dimensional emotion recognition based on the fusion of facial and vocal modalities extracted from videos, where complex spatiotemporal relationships may be captured. Most of the existing fusion techniques rely on recurrent networks or conventional attention mechanisms that do not effectively leverage the complimentary nature of audio-visual (A-V) modalities. We introduce a cross-attentional fusion approach to extract the salient features across A-V modalities, allowing for accurate prediction of continuous values of valence and arousal. Our new cross-attentional A-V fusion model efficiently leverages the inter-modal relationships. In particular, it computes cross-attention weights to focus on the more contributive features across individual modalities, and thereby combine contributive feature representations, which are then fed to fully connected layers for the prediction of valence and arousal. The effectiveness of the proposed approach is validated experimentally on videos from the RECOLA and Fatigue (private) data-sets. Results indicate that our cross-attentional A-V fusion model is a cost-effective approach that outperforms state-of-the-art fusion approaches. Code is available: \url{https://github.com/praveena2j/Cross-Attentional-AV-Fusion}
URLs: https://github.com/praveena2j/Cross-Attentional-AV-Fusion
Authors: Xu Song, Yongbiao Xiao, Hui Li, Xiao-Jun Wu, Jun Sun, Vasile Palade
Abstract: The fusion of visible light and infrared images has garnered significant attention in the field of imaging due to its pivotal role in various applications, including surveillance, remote sensing, and medical imaging. Therefore, this paper introduces a novel fusion framework using Res2Net architecture, capturing features across diverse receptive fields and scales for effective extraction of global and local features. Our methodology is structured into three fundamental components: the first part involves the Res2Net-based encoder, followed by the second part, which encompasses the fusion layer, and finally, the third part, which comprises the decoder. The encoder based on Res2Net is utilized for extracting multi-scale features from the input image. Simultaneously, with a single image as input, we introduce a pioneering training strategy tailored for a Res2Net-based encoder. We further enhance the fusion process with a novel strategy based on the attention model, ensuring precise reconstruction by the decoder for the fused image. Experimental results unequivocally showcase our method's unparalleled fusion performance, surpassing existing techniques, as evidenced by rigorous subjective and objective evaluations.
Authors: R. Gnana Praveen, Wheidima Carneiro de Melo, Nasib Ullah, Haseeb Aslam, Osama Zeeshan, Th\'eo Denorme, Marco Pedersoli, Alessandro Koerich, Simon Bacon, Patrick Cardinal, Eric Granger
Abstract: Multimodal emotion recognition has recently gained much attention since it can leverage diverse and complementary relationships over multiple modalities (e.g., audio, visual, biosignals, etc.), and can provide some robustness to noisy modalities. Most state-of-the-art methods for audio-visual (A-V) fusion rely on recurrent networks or conventional attention mechanisms that do not effectively leverage the complementary nature of A-V modalities. In this paper, we focus on dimensional emotion recognition based on the fusion of facial and vocal modalities extracted from videos. Specifically, we propose a joint cross-attention model that relies on the complementary relationships to extract the salient features across A-V modalities, allowing for accurate prediction of continuous values of valence and arousal. The proposed fusion model efficiently leverages the inter-modal relationships, while reducing the heterogeneity between the features. In particular, it computes the cross-attention weights based on correlation between the combined feature representation and individual modalities. By deploying the combined A-V feature representation into the cross-attention module, the performance of our fusion module improves significantly over the vanilla cross-attention module. Experimental results on validation-set videos from the AffWild2 dataset indicate that our proposed A-V fusion model provides a cost-effective solution that can outperform state-of-the-art approaches. The code is available on GitHub: https://github.com/praveena2j/JointCrossAttentional-AV-Fusion.
URLs: https://github.com/praveena2j/JointCrossAttentional-AV-Fusion.
Authors: Yifan Zhang, Qijian Zhang, Zhiyu Zhu, Junhui Hou, Yixuan Yuan
Abstract: The inherent ambiguity in ground-truth annotations of 3D bounding boxes, caused by occlusions, signal missing, or manual annotation errors, can confuse deep 3D object detectors during training, thus deteriorating detection accuracy. However, existing methods overlook such issues to some extent and treat the labels as deterministic. In this paper, we formulate the label uncertainty problem as the diversity of potentially plausible bounding boxes of objects. Then, we propose GLENet, a generative framework adapted from conditional variational autoencoders, to model the one-to-many relationship between a typical 3D object and its potential ground-truth bounding boxes with latent variables. The label uncertainty generated by GLENet is a plug-and-play module and can be conveniently integrated into existing deep 3D detectors to build probabilistic detectors and supervise the learning of the localization uncertainty. Besides, we propose an uncertainty-aware quality estimator architecture in probabilistic detectors to guide the training of the IoU-branch with predicted localization uncertainty. We incorporate the proposed methods into various popular base 3D detectors and demonstrate significant and consistent performance gains on both KITTI and Waymo benchmark datasets. Especially, the proposed GLENet-VR outperforms all published LiDAR-based approaches by a large margin and achieves the top rank among single-modal methods on the challenging KITTI test set. The source code and pre-trained models are publicly available at \url{https://github.com/Eaphan/GLENet}.
Authors: Ranjana Roy Chowdhury, Deepti R. Bathula
Abstract: Prototypical network (PN) is a simple yet effective few shot learning strategy. It is a metric-based meta-learning technique where classification is performed by computing Euclidean distances to prototypical representations of each class. Conventional PN attributes equal importance to all samples and generates prototypes by simply averaging the support sample embeddings belonging to each class. In this work, we propose a novel version of PN that attributes weights to support samples corresponding to their influence on the support sample distribution. Influence weights of samples are calculated based on maximum mean discrepancy (MMD) between the mean embeddings of sample distributions including and excluding the sample. Further, the influence factor of a sample is measured using MMD based on the shift in the distribution in the absence of that sample.
Authors: Adrien Doerig, Tim C Kietzmann, Emily Allen, Yihan Wu, Thomas Naselaris, Kendrick Kay, Ian Charest
Abstract: The human brain extracts complex information from visual inputs, including objects, their spatial and semantic interrelations, and their interactions with the environment. However, a quantitative approach for studying this information remains elusive. Here, we test whether the contextual information encoded in large language models (LLMs) is beneficial for modelling the complex visual information extracted by the brain from natural scenes. We show that LLM embeddings of scene captions successfully characterise brain activity evoked by viewing the natural scenes. This mapping captures selectivities of different brain areas, and is sufficiently robust that accurate scene captions can be reconstructed from brain activity. Using carefully controlled model comparisons, we then proceed to show that the accuracy with which LLM representations match brain representations derives from the ability of LLMs to integrate complex information contained in scene captions beyond that conveyed by individual words. Finally, we train deep neural network models to transform image inputs into LLM representations. Remarkably, these networks learn representations that are better aligned with brain representations than a large number of state-of-the-art alternative models, despite being trained on orders-of-magnitude less data. Overall, our results suggest that LLM embeddings of scene captions provide a representational format that accounts for complex information extracted by the brain from visual inputs.
Authors: Pavol Harar, Dennis Elbr\"achter, Monika D\"orfler, Kory D. Johnson
Abstract: We present an algorithm and package, Redistributor, which forces a collection of scalar samples to follow a desired distribution. When given independent and identically distributed samples of some random variable $S$ and the continuous cumulative distribution function of some desired target $T$, it provably produces a consistent estimator of the transformation $R$ which satisfies $R(S)=T$ in distribution. As the distribution of $S$ or $T$ may be unknown, we also include algorithms for efficiently estimating these distributions from samples. This allows for various interesting use cases in image processing, where Redistributor serves as a remarkably simple and easy-to-use tool that is capable of producing visually appealing results. For color correction it outperforms other model-based methods and excels in achieving photorealistic style transfer, surpassing deep learning methods in content preservation. The package is implemented in Python and is optimized to efficiently handle large datasets, making it also suitable as a preprocessing step in machine learning. The source code is available at https://github.com/paloha/redistributor.
Authors: Marziyeh Bamdad, Davide Scaramuzza, Alireza Darvishy
Abstract: In recent decades, several assistive technologies have been developed to improve the ability of blind and visually impaired individuals to navigate independently and safely. At the same time, simultaneous localization and mapping (SLAM) techniques have become sufficiently robust and efficient to be adopted in developing these assistive technologies. We present the first systematic literature review of 54 recent studies on SLAM-based solutions for blind and visually impaired people, focusing on literature published from 2017 onward. This review explores various localization and mapping techniques employed in this context. We systematically identified and categorized diverse SLAM approaches and analyzed their localization and mapping techniques, sensor types, computing resources, and machine-learning methods. We discuss the advantages and limitations of these techniques for blind and visually impaired navigation. Moreover, we examine the major challenges described across studies, including practical considerations that affect usability and adoption. Our analysis also evaluates the effectiveness of these SLAM-based solutions in real-world scenarios and user satisfaction, providing insights into their practical impact on BVI mobility. The insights derived from this review identify critical gaps and opportunities for future research activities, particularly in addressing the challenges presented by dynamic and complex environments. We explain how SLAM technology offers the potential to improve the ability of visually impaired individuals to navigate effectively. Finally, we present future opportunities and challenges in this domain.
Authors: Shashank Agnihotri, Steffen Jung, Margret Keuper
Abstract: While neural networks allow highly accurate predictions in many tasks, their lack of robustness towards even slight input perturbations often hampers their deployment. Adversarial attacks such as the seminal projected gradient descent (PGD) offer an effective means to evaluate a model's robustness and dedicated solutions have been proposed for attacks on semantic segmentation or optical flow estimation. While they attempt to increase the attack's efficiency, a further objective is to balance its effect, so that it acts on the entire image domain instead of isolated point-wise predictions. This often comes at the cost of optimization stability and thus efficiency. Here, we propose CosPGD, an attack that encourages more balanced errors over the entire image domain while increasing the attack's overall efficiency. To this end, CosPGD leverages a simple alignment score computed from any pixel-wise prediction and its target to scale the loss in a smooth and fully differentiable way. It leads to efficient evaluations of a model's robustness for semantic segmentation as well as regression models (such as optical flow, disparity estimation, or image restoration), and it allows it to outperform the previous SotA attack on semantic segmentation. We provide code for the CosPGD algorithm and example usage at https://github.com/shashankskagnihotri/cospgd.
Authors: Shen Yan, Xiaoya Cheng, Yuxiang Liu, Juelin Zhu, Rouwan Wu, Yu Liu, Maojun Zhang
Abstract: Despite the significant progress in 6-DoF visual localization, researchers are mostly driven by ground-level benchmarks. Compared with aerial oblique photography, ground-level map collection lacks scalability and complete coverage. In this work, we propose to go beyond the traditional ground-level setting and exploit the cross-view localization from aerial to ground. We solve this problem by formulating camera pose estimation as an iterative render-and-compare pipeline and enhancing the robustness through augmenting seeds from noisy initial priors. As no public dataset exists for the studied problem, we collect a new dataset that provides a variety of cross-view images from smartphones and drones and develop a semi-automatic system to acquire ground-truth poses for query images. We benchmark our method as well as several state-of-the-art baselines and demonstrate that our method outperforms other approaches by a large margin.
Authors: Jian Shi, Pengyi Zhang, Ni Zhang, Hakim Ghazzai, Peter Wonka
Abstract: Medical imaging often contains critical fine-grained features, such as tumors or hemorrhages, crucial for diagnosis yet potentially too subtle for detection with conventional methods. In this paper, we introduce \textit{DIA}, dissolving is amplifying. DIA is a fine-grained anomaly detection framework for medical images. First, we introduce \textit{dissolving transformations}. We employ diffusion with a generative diffusion model as a dedicated feature-aware denoiser. Applying diffusion to medical images in a certain manner can remove or diminish fine-grained discriminative features. Second, we introduce an \textit{amplifying framework} based on contrastive learning to learn a semantically meaningful representation of medical images in a self-supervised manner, with a focus on fine-grained features. The amplifying framework contrasts additional pairs of images with and without dissolving transformations applied and thereby emphasizes the dissolved fine-grained features. DIA significantly improves the medical anomaly detection performance with around 18.40\% AUC boost against the baseline method and achieves an overall SOTA against other benchmark methods. Our code is available at \url{https://github.com/shijianjian/DIA.git}.
Authors: Ruiqi Wang, Akshay Gadi Patil, Fenggen Yu, Hao Zhang
Abstract: We introduce the first active learning (AL) model for high-accuracy instance segmentation of moveable parts from RGB images of real indoor scenes. Specifically, our goal is to obtain fully validated segmentation results by humans while minimizing manual effort. To this end, we employ a transformer that utilizes a masked-attention mechanism to supervise the active segmentation. To enhance the network tailored to moveable parts, we introduce a coarse-to-fine AL approach which first uses an object-aware masked attention and then a pose-aware one, leveraging the hierarchical nature of the problem and a correlation between moveable parts and object poses and interaction directions. When applying our AL model to 2,000 real images, we obtain fully validated moveable part segmentations with semantic labels, by only needing to manually annotate 11.45% of the images. This translates to significant (60%) time saving over manual effort required by the best non-AL model to attain the same segmentation accuracy. At last, we contribute a dataset of 2,550 real images with annotated moveable parts, demonstrating its superior quality and diversity over the best alternatives.
Authors: Patrick M{\o}ller Jensen, Vedrana Andersen Dahl, Carsten Gundlach, Rebecca Engberg, Hans Martin Kjer, Anders Bjorholm Dahl
Abstract: Domain shift significantly influences the performance of deep learning algorithms, particularly for object detection within volumetric 3D images. Annotated training data is essential for deep learning-based object detection. However, annotating densely packed objects is time-consuming and costly. Instead, we suggest training models on individually scanned objects, causing a domain shift between training and detection data. To address this challenge, we introduce the BugNIST dataset, comprising 9154 micro-CT volumes of 12 bug types and 388 volumes of tightly packed bug mixtures. This dataset is characterized by having objects with the same appearance in the source and target domains, which is uncommon for other benchmark datasets for domain shift. During training, individual bug volumes labeled by class are utilized, while testing employs mixtures with center point annotations and bug type labels. Together with the dataset, we provide a baseline detection analysis, with the aim of advancing the field of 3D object detection methods.
Authors: Astha Verma, A V Subramanyam, Siddhesh Bangar, Naman Lal, Rajiv Ratn Shah, Shin'ichi Satoh
Abstract: Certified defense methods against adversarial perturbations have been recently investigated in the black-box setting with a zeroth-order (ZO) perspective. However, these methods suffer from high model variance with low performance on high-dimensional datasets due to the ineffective design of the denoiser and are limited in their utilization of ZO techniques. To this end, we propose a certified ZO preprocessing technique for removing adversarial perturbations from the attacked image in the black-box setting using only model queries. We propose a robust UNet denoiser (RDUNet) that ensures the robustness of black-box models trained on high-dimensional datasets. We propose a novel black-box denoised smoothing (DS) defense mechanism, ZO-RUDS, by prepending our RDUNet to the black-box model, ensuring black-box defense. We further propose ZO-AE-RUDS in which RDUNet followed by autoencoder (AE) is prepended to the black-box model. We perform extensive experiments on four classification datasets, CIFAR-10, CIFAR-10, Tiny Imagenet, STL-10, and the MNIST dataset for image reconstruction tasks. Our proposed defense methods ZO-RUDS and ZO-AE-RUDS beat SOTA with a huge margin of $35\%$ and $9\%$, for low dimensional (CIFAR-10) and with a margin of $20.61\%$ and $23.51\%$ for high-dimensional (STL-10) datasets, respectively.
Authors: Jiawei Qin, Takuru Shimoyama, Xucong Zhang, Yusuke Sugano
Abstract: Along with the recent development of deep neural networks, appearance-based gaze estimation has succeeded considerably when training and testing within the same domain. Compared to the within-domain task, the variance of different domains makes the cross-domain performance drop severely, preventing gaze estimation deployment in real-world applications. Among all the factors, ranges of head pose and gaze are believed to play significant roles in the final performance of gaze estimation, while collecting large ranges of data is expensive. This work proposes an effective model training pipeline consisting of a training data synthesis and a gaze estimation model for unsupervised domain adaptation. The proposed data synthesis leverages the single-image 3D reconstruction to expand the range of the head poses from the source domain without requiring a 3D facial shape dataset. To bridge the inevitable gap between synthetic and real images, we further propose an unsupervised domain adaptation method suitable for synthetic full-face data. We propose a disentangling autoencoder network to separate gaze-related features and introduce background augmentation consistency loss to utilize the characteristics of the synthetic source domain. Through comprehensive experiments, it shows that the model using only our synthetic training data can perform comparably to real data extended with a large label range. Our proposed domain adaptation approach further improves the performance on multiple target domains. The code and data will be available at https://github.com/ut-vision/AdaptiveGaze.
Authors: Shady Abu-Hussein, Raja Giryes
Abstract: Denoising Diffusion Probabilistic Models (DDPM) have recently gained significant attention. DDPMs compose a Markovian process that begins in the data domain and gradually adds noise until reaching pure white noise. DDPMs generate high-quality samples from complex data distributions by defining an inverse process and training a deep neural network to learn this mapping. However, these models are inefficient because they require many diffusion steps to produce aesthetically pleasing samples. Additionally, unlike generative adversarial networks (GANs), the latent space of diffusion models is less interpretable. In this work, we propose to generalize the denoising diffusion process into an Upsampling Diffusion Probabilistic Model (UDPM). In the forward process, we reduce the latent variable dimension through downsampling, followed by the traditional noise perturbation. As a result, the reverse process gradually denoises and upsamples the latent variable to produce a sample from the data distribution. We formalize the Markovian diffusion processes of UDPM and demonstrate its generation capabilities on the popular FFHQ, AFHQv2, and CIFAR10 datasets. UDPM generates images with as few as three network evaluations, whose overall computational cost is less than a single DDPM or EDM step, while achieving an FID score of 6.86. This surpasses current state-of-the-art efficient diffusion models that use a single denoising step for sampling. Additionally, UDPM offers an interpretable and interpolable latent space, which gives it an advantage over traditional DDPMs. Our code is available online: \url{https://github.com/shadyabh/UDPM/}
Authors: Hanxi Li, Jingqi Wu, Lin Yuanbo Wu, Hao Chen, Mingwen Wang, Chunhua Shen
Abstract: Recent advancements in industrial Anomaly Detection (AD) have shown that incorporating a few anomalous samples during training can significantly boost accuracy. However, this performance improvement comes at a high cost: extensive annotation efforts, which are often impractical in real-world applications. In this work, we propose a novel framework called ``Weakly-supervised RESidual Transformer`` (WeakREST), which aims to achieve high AD accuracy while minimizing the need for extensive annotations. First, we reformulate the pixel-wise anomaly localization task into a block-wise classification problem. By shifting the focus to block-wise level, we can drastically reduce the amount of required annotations without compromising on the accuracy of anomaly detection Secondly, we design a residual-based transformer model, termed ``Positional Fast Anomaly Residuals`` (PosFAR), to classify the image blocks in real time. We further propose to label the anomalous regions using only bounding boxes or image tags as weaker labels, leading to a semi-supervised learning setting. On the benchmark dataset MVTec-AD, our proposed WeakREST framework achieves a remarkable Average Precision (AP) of 83.0%, significantly outperforming the previous best result of 75.8% in the unsupervised setting. In the supervised AD setting, WeakREST further improves performance, attaining an AP of 87.6% compared to the previous best of 78.6%. Notably, even when utilizing weaker labels based on bounding boxes, WeakREST surpasses recent leading methods that rely on pixel-wise supervision, achieving an AP of 87.1% against the prior best of 78.6% on MVTec-AD. This precision advantage is also consistently observed on other well-known AD datasets, such as BTAD and KSDD2.
Authors: Hao Guan, Pew-Thian Yap, Andrea Bozoki, Mingxia Liu
Abstract: Machine learning in medical imaging often faces a fundamental dilemma, namely, the small sample size problem. Many recent studies suggest using multi-domain data pooled from different acquisition sites/centers to improve statistical power. However, medical images from different sites cannot be easily shared to build large datasets for model training due to privacy protection reasons. As a promising solution, federated learning, which enables collaborative training of machine learning models based on data from different sites without cross-site data sharing, has attracted considerable attention recently. In this paper, we conduct a comprehensive survey of the recent development of federated learning methods in medical image analysis. In this survey, we first introduce the background knowledge of federated learning for dealing with privacy protection and collaborative learning issues in medical imaging. We then present a comprehensive review of recent advances in federated learning methods for medical image analysis. Specifically, existing methods are categorized based on three critical aspects of a federated learning system, including client end, server end, and communication techniques. In each category, we summarize the existing federated learning methods according to specific research problems in medical image analysis and also provide insights into the motivations of different approaches. In addition, we provide a review of existing benchmark medical imaging datasets and software platforms for current federated learning research. We also conduct an experimental study to empirically evaluate typical federated learning methods for medical image analysis. This survey can help to better understand the current research status, challenges, and potential research opportunities in this promising research field.
Authors: Amirhossein Aminimehr, Pouya Khani, Amirali Molaei, Amirmohammad Kazemeini, Erik Cambria
Abstract: The field of Explainable Artificial Intelligence (XAI) aims to improve the interpretability of black-box machine learning models. Building a heatmap based on the importance value of input features is a popular method for explaining the underlying functions of such models in producing their predictions. Heatmaps are almost understandable to humans, yet they are not without flaws. Non-expert users, for example, may not fully understand the logic of heatmaps (the logic in which relevant pixels to the model's prediction are highlighted with different intensities or colors). Additionally, objects and regions of the input image that are relevant to the model prediction are frequently not entirely differentiated by heatmaps. In this paper, we propose a framework called TbExplain that employs XAI techniques and a pre-trained object detector to present text-based explanations of scene classification models. Moreover, TbExplain incorporates a novel method to correct predictions and textually explain them based on the statistics of objects in the input image when the initial prediction is unreliable. To assess the trustworthiness and validity of the text-based explanations, we conducted a qualitative experiment, and the findings indicated that these explanations are sufficiently reliable. Furthermore, our quantitative and qualitative experiments on TbExplain with scene classification datasets reveal an improvement in classification accuracy over ResNet variants.
Authors: Ioannis Maniadis Metaxas, Adrian Bulat, Ioannis Patras, Brais Martinez, Georgios Tzimiropoulos
Abstract: The unsupervised pretraining of object detectors has recently become a key component of object detector training, as it leads to improved performance and faster convergence during the supervised fine-tuning stage. Existing unsupervised pretraining methods, however, typically rely on low-level information to define proposals that are used to train the detector. Furthermore, in the absence of class labels for these proposals, an auxiliary loss is used to add high-level semantics. This results in complex pipelines and a task gap between the pretraining and the downstream task. We propose a framework that mitigates this issue and consists of three simple yet key ingredients: (i) richer initial proposals that do encode high-level semantics, (ii) class pseudo-labeling through clustering, that enables pretraining using a standard object detection training pipeline, (iii) self-training to iteratively improve and enrich the object proposals. Once the pretraining and downstream tasks are aligned, a simple detection pipeline without further bells and whistles can be directly used for pretraining and, in fact, results in state-of-the-art performance on both the full and low data regimes, across detector architectures and datasets, by significant margins. We further show that our pretraining strategy is also capable of pretraining from scratch (including the backbone) and works on complex images like COCO, paving the path for unsupervised representation learning using object detection directly as a pretext task.
Authors: Michail Mamalakis, Heloise de Vareilles, Atheer AI-Manea, Samantha C. Mitchell, Ingrid Arartz, Lynn Egeland Morch-Johnsen, Jane Garrison, Jon Simons, Pietro Lio, John Suckling, Graham Murray
Abstract: The significant features identified in a representative subset of the dataset during the learning process of an artificial intelligence model are referred to as a 'global' explanation. Three-dimensional (3D) global explanations are crucial in neuroimaging where a complex representational space demands more than basic two-dimensional interpretations. Curently, studies in the literature lack accurate, low-complexity, and 3D global explanations in neuroimaging and beyond. To fill this gap, we develop a novel explainable artificial intelligence (XAI) 3D-Framework that provides robust, faithful, and low-complexity global explanations. We evaluated our framework on various 3D deep learning networks trained, validated, and tested on a well-annotated cohort of 596 MRI images. The focus of detection was on the presence or absence of the paracingulate sulcus, a highly variable feature of brain topology associated with symptoms of psychosis. Our proposed 3D-Framework outperformed traditional XAI methods in terms of faithfulness for global explanations. As a result, these explanations uncovered new patterns that not only enhance the credibility and reliability of the training process but also reveal the broader developmental landscape of the human cortex. Our XAI 3D-Framework proposes for the first time, a way to utilize global explanations to discover the context in which detection of specific features are embedded, opening our understanding of normative brain development and atypical trajectories that can lead to the emergence of mental illness.
Authors: Xian Lin, Yangyang Xiang, Li Yu, Zengqiang Yan
Abstract: End-to-end medical image segmentation is of great value for computer-aided diagnosis dominated by task-specific models, usually suffering from poor generalization. With recent breakthroughs brought by the segment anything model (SAM) for universal image segmentation, extensive efforts have been made to adapt SAM for medical imaging but still encounter two major issues: 1) severe performance degradation and limited generalization without proper adaptation, and 2) semi-automatic segmentation relying on accurate manual prompts for interaction. In this work, we propose SAMUS as a universal model tailored for ultrasound image segmentation and further enable it to work in an end-to-end manner denoted as AutoSAMUS. Specifically, in SAMUS, a parallel CNN branch is introduced to supplement local information through cross-branch attention, and a feature adapter and a position adapter are jointly used to adapt SAM from natural to ultrasound domains while reducing training complexity. AutoSAMUS is realized by introducing an auto prompt generator (APG) to replace the manual prompt encoder of SAMUS to automatically generate prompt embeddings. A comprehensive ultrasound dataset, comprising about 30k images and 69k masks and covering six object categories, is collected for verification. Extensive comparison experiments demonstrate the superiority of SAMUS and AutoSAMUS against the state-of-the-art task-specific and SAM-based foundation models. We believe the auto-prompted SAM-based model has the potential to become a new paradigm for end-to-end medical image segmentation and deserves more exploration. Code and data are available at https://github.com/xianlin7/SAMUS.
Authors: Yimeng Zhang, Jinghan Jia, Xin Chen, Aochuan Chen, Yihua Zhang, Jiancheng Liu, Ke Ding, Sijia Liu
Abstract: The recent advances in diffusion models (DMs) have revolutionized the generation of realistic and complex images. However, these models also introduce potential safety hazards, such as producing harmful content and infringing data copyrights. Despite the development of safety-driven unlearning techniques to counteract these challenges, doubts about their efficacy persist. To tackle this issue, we introduce an evaluation framework that leverages adversarial prompts to discern the trustworthiness of these safety-driven DMs after they have undergone the process of unlearning harmful concepts. Specifically, we investigated the adversarial robustness of DMs, assessed by adversarial prompts, when eliminating unwanted concepts, styles, and objects. We develop an effective and efficient adversarial prompt generation approach for DMs, termed UnlearnDiffAtk. This method capitalizes on the intrinsic classification abilities of DMs to simplify the creation of adversarial prompts, thereby eliminating the need for auxiliary classification or diffusion models. Through extensive benchmarking, we evaluate the robustness of widely-used safety-driven unlearned DMs (i.e., DMs after unlearning undesirable concepts, styles, or objects) across a variety of tasks. Our results demonstrate the effectiveness and efficiency merits of UnlearnDiffAtk over the state-of-the-art adversarial prompt generation method and reveal the lack of robustness of current safetydriven unlearning techniques when applied to DMs. Codes are available at https://github.com/OPTML-Group/Diffusion-MU-Attack. WARNING: There exist AI generations that may be offensive in nature.
Authors: Siao Tang, Xin Wang, Hong Chen, Chaoyu Guan, Zewen Wu, Yansong Tang, Wenwu Zhu
Abstract: High computational overhead is a troublesome problem for diffusion models. Recent studies have leveraged post-training quantization (PTQ) to compress diffusion models. However, most of them only focus on unconditional models, leaving the quantization of widely-used pretrained text-to-image models, e.g., Stable Diffusion, largely unexplored. In this paper, we propose a novel post-training quantization method PCR (Progressive Calibration and Relaxing) for text-to-image diffusion models, which consists of a progressive calibration strategy that considers the accumulated quantization error across timesteps, and an activation relaxing strategy that improves the performance with negligible cost. Additionally, we demonstrate the previous metrics for text-to-image diffusion model quantization are not accurate due to the distribution gap. To tackle the problem, we propose a novel QDiffBench benchmark, which utilizes data in the same domain for more accurate evaluation. Besides, QDiffBench also considers the generalization performance of the quantized model outside the calibration dataset. Extensive experiments on Stable Diffusion and Stable Diffusion XL demonstrate the superiority of our method and benchmark. Moreover, we are the first to achieve quantization for Stable Diffusion XL while maintaining the performance.
Authors: Yimeng Shan, Xuerui Qiu, Rui-jie Zhu, Jason K. Eshraghian, Malu Zhang, Haicheng Qu
Abstract: Spiking Neural Networks (SNNs) have garnered substantial attention in brain-like computing for their biological fidelity and the capacity to execute energy-efficient spike-driven operations. As the demand for heightened performance in SNNs surges, the trend towards training deeper networks becomes imperative, while residual learning stands as a pivotal method for training deep neural networks. In our investigation, we identified that the SEW-ResNet, a prominent representative of deep residual spiking neural networks, incorporates non-event-driven operations. To rectify this, we propose a novel training paradigm that first accumulates a large amount of redundant information through OR Residual Connection (ORRC), and then filters out the redundant information using the Synergistic Attention (SynA) module, which promotes feature extraction in the backbone while suppressing the influence of noise and useless features in the shortcuts. When integrating SynA into the network, we observed the phenomenon of "natural pruning", where after training, some or all of the shortcuts in the network naturally drop out without affecting the model's classification accuracy. This significantly reduces computational overhead and makes it more suitable for deployment on edge devices. Experimental results on various public datasets confirmed that the SynA-ResNet achieved single-sample classification with as little as 0.8 spikes per neuron. Moreover, when compared to other residual SNN models, it exhibited higher accuracy and up to a 28-fold reduction in energy consumption.
Authors: Bardia Khosravi, Frank Li, Theo Dapamede, Pouria Rouzrokh, Cooper U. Gamble, Hari M. Trivedi, Cody C. Wyles, Andrew B. Sellergren, Saptarshi Purkayastha, Bradley J. Erickson, Judy W. Gichoya
Abstract: Chest X-rays (CXR) are essential for diagnosing a variety of conditions, but when used on new populations, model generalizability issues limit their efficacy. Generative AI, particularly denoising diffusion probabilistic models (DDPMs), offers a promising approach to generating synthetic images, enhancing dataset diversity. This study investigates the impact of synthetic data supplementation on the performance and generalizability of medical imaging research. The study employed DDPMs to create synthetic CXRs conditioned on demographic and pathological characteristics from the CheXpert dataset. These synthetic images were used to supplement training datasets for pathology classifiers, with the aim of improving their performance. The evaluation involved three datasets (CheXpert, MIMIC-CXR, and Emory Chest X-ray) and various experiments, including supplementing real data with synthetic data, training with purely synthetic data, and mixing synthetic data with external datasets. Performance was assessed using the area under the receiver operating curve (AUROC). Adding synthetic data to real datasets resulted in a notable increase in AUROC values (up to 0.02 in internal and external test sets with 1000% supplementation, p-value less than 0.01 in all instances). When classifiers were trained exclusively on synthetic data, they achieved performance levels comparable to those trained on real data with 200%-300% data supplementation. The combination of real and synthetic data from different sources demonstrated enhanced model generalizability, increasing model AUROC from 0.76 to 0.80 on the internal test set (p-value less than 0.01). In conclusion, synthetic data supplementation significantly improves the performance and generalizability of pathology classifiers in medical imaging.
Authors: Yankun Xu, Junzhe Wang, Yun-Hsuan Chen, Jie Yang, Wenjie Ming, Shuang Wang, Mohamad Sawan
Abstract: An accurate and efficient epileptic seizure onset detection can significantly benefit patients. Traditional diagnostic methods, primarily relying on electroencephalograms (EEGs), often result in cumbersome and non-portable solutions, making continuous patient monitoring challenging. The video-based seizure detection system is expected to free patients from the constraints of scalp or implanted EEG devices and enable remote monitoring in residential settings. Previous video-based methods neither enable all-day monitoring nor provide short detection latency due to insufficient resources and ineffective patient action recognition techniques. Additionally, skeleton-based action recognition approaches remain limitations in identifying subtle seizure-related actions. To address these challenges, we propose a novel Video-based Seizure detection model via a skeleton-based spatiotemporal Vision Graph neural network (VSViG) for its efficient, accurate and timely purpose in real-time scenarios. Our experimental results indicate VSViG outperforms previous state-of-the-art action recognition models on our collected patients' video data with higher accuracy (5.9% error), lower FLOPs (0.4G), and smaller model size (1.4M). Furthermore, by integrating a decision-making rule that combines output probabilities and an accumulative function, we achieve a 5.1 s detection latency after EEG onset, a 13.1 s detection advance before clinical onset, and a zero false detection rate. The project homepage is available at: https://github.com/xuyankun/VSViG/
Authors: Shuangrui Ding, Rui Qian, Haohang Xu, Dahua Lin, Hongkai Xiong
Abstract: In this paper, we propose a simple yet effective approach for self-supervised video object segmentation (VOS). Our key insight is that the inherent structural dependencies present in DINO-pretrained Transformers can be leveraged to establish robust spatio-temporal correspondences in videos. Furthermore, simple clustering on this correspondence cue is sufficient to yield competitive segmentation results. Previous self-supervised VOS techniques majorly resort to auxiliary modalities or utilize iterative slot attention to assist in object discovery, which restricts their general applicability and imposes higher computational requirements. To deal with these challenges, we develop a simplified architecture that capitalizes on the emerging objectness from DINO-pretrained Transformers, bypassing the need for additional modalities or slot attention. Specifically, we first introduce a single spatio-temporal Transformer block to process the frame-wise DINO features and establish spatio-temporal dependencies in the form of self-attention. Subsequently, utilizing these attention maps, we implement hierarchical clustering to generate object segmentation masks. To train the spatio-temporal block in a fully self-supervised manner, we employ semantic and dynamic motion consistency coupled with entropy normalization. Our method demonstrates state-of-the-art performance across multiple unsupervised VOS benchmarks and particularly excels in complex real-world multi-object video segmentation tasks such as DAVIS-17-Unsupervised and YouTube-VIS-19. The code and model checkpoints will be released at https://github.com/shvdiwnkozbw/SSL-UVOS.
Authors: Chinmay Savadikar, Xi Song, Tianfu Wu
Abstract: We present Generative Interpretable Fine-Tuning (GIFT) for parameter-efficient fine-tuning of pretrained Transformer backbones, which can be formulated as a simple factorized matrix multiplication in the parameter space or equivalently in the activation/representation space, and thus embraces built-in interpretability. For a layer with weights $\omega\in \mathbb{R}^{d_{out}\times d_{in}}$, our proposed GIFT learns the fine-tuned weights $\hat{\omega}$ directly from $\omega$ as $\hat{\omega}=\omega\cdot (\mathbb{I}+\phi_{d_{in}\times r}\cdot\psi_{r\times d_{in}})$. $\Theta=(\phi, \psi)$ are the learnable parameters of the two linear layers. $\Theta$ can be shared by all layers selected for fine-tuning (e.g., all the Query and Value layers), or can be layer-type specific (e.g., different $\Theta$'s used for Query and Value), resulting in significantly fewer trainable parameters compared to layer-specific Low-Rank Adaptation (LoRA). We perform comprehensive evaluations on natural language tasks (commonsense and arithmetic reasoning, instruction tuning, and sequence classification), and fine-grained visual classification tasks. We obtain the best performance and parameter efficiency among baselines on commonsense reasoning, instruction tuning and visual recognition benchmarks. Compared to LoRA, we obtain 5.9% absolute increase in average accuracy with 53.8 times reduction of parameters on Commonsense170k using Llama-3 (8B), and 5.4% absolute increase in the win rate with 4 times reduction of parameters using Llama-2 (7B) during instruction tuning. Our GIFT also obtains a slightly higher win rate on instruction tuning than GPT 3.5 (Turbo 1106). We show the output of the first linear layer (i.e., $\omega\cdot \phi$) is surprisingly interpretable, which can play the role of a token-clustering head as a by-product to localize meaningful objects/parts in images for computer vision tasks.
Authors: Mingqiao Ye, Martin Danelljan, Fisher Yu, Lei Ke
Abstract: The recent Gaussian Splatting achieves high-quality and real-time novel-view synthesis of the 3D scenes. However, it is solely concentrated on the appearance and geometry modeling, while lacking in fine-grained object-level scene understanding. To address this issue, we propose Gaussian Grouping, which extends Gaussian Splatting to jointly reconstruct and segment anything in open-world 3D scenes. We augment each Gaussian with a compact Identity Encoding, allowing the Gaussians to be grouped according to their object instance or stuff membership in the 3D scene. Instead of resorting to expensive 3D labels, we supervise the Identity Encodings during the differentiable rendering by leveraging the 2D mask predictions by Segment Anything Model (SAM), along with introduced 3D spatial consistency regularization. Compared to the implicit NeRF representation, we show that the discrete and grouped 3D Gaussians can reconstruct, segment and edit anything in 3D with high visual quality, fine granularity and efficiency. Based on Gaussian Grouping, we further propose a local Gaussian Editing scheme, which shows efficacy in versatile scene editing applications, including 3D object removal, inpainting, colorization, style transfer and scene recomposition. Our code and models are at https://github.com/lkeab/gaussian-grouping.
Authors: Alex Costanzino, Pierluigi Zama Ramirez, Giuseppe Lisanti, Luigi Di Stefano
Abstract: The paper explores the industrial multimodal Anomaly Detection (AD) task, which exploits point clouds and RGB images to localize anomalies. We introduce a novel light and fast framework that learns to map features from one modality to the other on nominal samples. At test time, anomalies are detected by pinpointing inconsistencies between observed and mapped features. Extensive experiments show that our approach achieves state-of-the-art detection and segmentation performance in both the standard and few-shot settings on the MVTec 3D-AD dataset while achieving faster inference and occupying less memory than previous multimodal AD methods. Moreover, we propose a layer-pruning technique to improve memory and time efficiency with a marginal sacrifice in performance.
Authors: Thuan Hoang Nguyen, Anh Tran
Abstract: Despite their ability to generate high-resolution and diverse images from text prompts, text-to-image diffusion models often suffer from slow iterative sampling processes. Model distillation is one of the most effective directions to accelerate these models. However, previous distillation methods fail to retain the generation quality while requiring a significant amount of images for training, either from real data or synthetically generated by the teacher model. In response to this limitation, we present a novel image-free distillation scheme named $\textbf{SwiftBrush}$. Drawing inspiration from text-to-3D synthesis, in which a 3D neural radiance field that aligns with the input prompt can be obtained from a 2D text-to-image diffusion prior via a specialized loss without the use of any 3D data ground-truth, our approach re-purposes that same loss for distilling a pretrained multi-step text-to-image model to a student network that can generate high-fidelity images with just a single inference step. In spite of its simplicity, our model stands as one of the first one-step text-to-image generators that can produce images of comparable quality to Stable Diffusion without reliance on any training image data. Remarkably, SwiftBrush achieves an FID score of $\textbf{16.67}$ and a CLIP score of $\textbf{0.29}$ on the COCO-30K benchmark, achieving competitive results or even substantially surpassing existing state-of-the-art distillation techniques.
Authors: Yuchen Yang, Yu Qiao, Xiao Sun
Abstract: Automatic estimation of 3D human pose from monocular RGB images is a challenging and unsolved problem in computer vision. In a supervised manner, approaches heavily rely on laborious annotations and present hampered generalization ability due to the limited diversity of 3D pose datasets. To address these challenges, we propose a unified framework that leverages mask as supervision for unsupervised 3D pose estimation. With general unsupervised segmentation algorithms, the proposed model employs skeleton and physique representations that exploit accurate pose information from coarse to fine. Compared with previous unsupervised approaches, we organize the human skeleton in a fully unsupervised way which enables the processing of annotation-free data and provides ready-to-use estimation results. Comprehensive experiments demonstrate our state-of-the-art pose estimation performance on Human3.6M and MPI-INF-3DHP datasets. Further experiments on in-the-wild datasets also illustrate the capability to access more data to boost our model. Code will be available at https://github.com/Charrrrrlie/Mask-as-Supervision.
Authors: Jing Li, Qiu-Feng Wang, Siyuan Wang, Rui Zhang, Kaizhu Huang, Erik Cambria
Abstract: Deciphering oracle bone scripts plays an important role in Chinese archaeology and philology. However, a significant challenge remains due to the scarcity of oracle character images. To overcome this issue, we propose Diff-Oracle, a novel approach based on diffusion models to generate a diverse range of controllable oracle characters. Unlike traditional diffusion models that operate primarily on text prompts, Diff-Oracle incorporates a style encoder that utilizes style reference images to control the generation style. This encoder extracts style prompts from existing oracle character images, where style details are converted into a text embedding format via a pretrained language-vision model. On the other hand, a content encoder is integrated within Diff-Oracle to capture specific content details from content reference images, ensuring that the generated characters accurately represent the intended glyphs. To effectively train Diff-Oracle, we pre-generate pixel-level paired oracle character images (i.e., style and content images) by an image-to-image translation model. Extensive qualitative and quantitative experiments are conducted on datasets Oracle-241 and OBC306. While significantly surpassing present generative methods in terms of image generation, Diff-Oracle substantially benefits downstream oracle character recognition, outperforming all existing SOTAs by a large margin. In particular, on the challenging OBC306 dataset, Diff-Oracle leads to an accuracy gain of 7.70% in the zero-shot setting and is able to recognize unseen oracle character images with the accuracy of 84.62%, achieving a new benchmark for deciphering oracle bone scripts.
Authors: Jiaxi Wang, Wenhui Hu, Xueyang Liu, Beihu Wu, Yuting Qiu, YingYing Cai
Abstract: Visual grounding aims to align visual information of specific regions of images with corresponding natural language expressions. Current visual grounding methods leverage pre-trained visual and language backbones independently to obtain visual features and linguistic features. Although these two types of features are then fused through elaborately designed networks, the heterogeneity of the features renders them unsuitable for multi-modal reasoning. This problem arises from the domain gap between the single-modal pre-training backbones used in current visual grounding methods, which can hardly be bridged by the traditional end-to-end training method. To alleviate this, our work proposes an Empowering Pre-trained Model for Visual Grounding (EpmVG) framework, which distills a multimodal pre-trained model to guide the visual grounding task. EpmVG relies on a novel cross-modal distillation mechanism that can effectively introduce the consistency information of images and texts from the pre-trained model, reducing the domain gap in the backbone networks, and thereby improving the performance of the model in the visual grounding task. Extensive experiments have been conducted on five conventionally used datasets, and the results demonstrate that our method achieves better performance than state-of-the-art methods.
Authors: Guikun Chen, Wenguan Wang
Abstract: 3D Gaussian splatting (GS) has recently emerged as a transformative technique in the realm of explicit radiance field and computer graphics. This innovative approach, characterized by the utilization of millions of learnable 3D Gaussians, represents a significant departure from mainstream neural radiance field approaches, which predominantly use implicit, coordinate-based models to map spatial coordinates to pixel values. 3D GS, with its explicit scene representation and differentiable rendering algorithm, not only promises real-time rendering capability but also introduces unprecedented levels of editability. This positions 3D GS as a potential game-changer for the next generation of 3D reconstruction and representation. In the present paper, we provide the first systematic overview of the recent developments and critical contributions in the domain of 3D GS. We begin with a detailed exploration of the underlying principles and the driving forces behind the emergence of 3D GS, laying the groundwork for understanding its significance. A focal point of our discussion is the practical applicability of 3D GS. By enabling unprecedented rendering speed, 3D GS opens up a plethora of applications, ranging from virtual reality to interactive media and beyond. This is complemented by a comparative analysis of leading 3D GS models, evaluated across various benchmark tasks to highlight their performance and practical utility. The survey concludes by identifying current challenges and suggesting potential avenues for future research in this domain. Through this survey, we aim to provide a valuable resource for both newcomers and seasoned researchers, fostering further exploration and advancement in applicable and explicit radiance field representation.
Authors: David Rivas-Villar, \'Alvaro S. Hervella, Jos\'e Rouco, Jorge Novo
Abstract: Retinal image registration is of utmost importance due to its wide applications in medical practice. In this context, we propose ConKeD, a novel deep learning approach to learn descriptors for retinal image registration. In contrast to current registration methods, our approach employs a novel multi-positive multi-negative contrastive learning strategy that enables the utilization of additional information from the available training samples. This makes it possible to learn high quality descriptors from limited training data. To train and evaluate ConKeD, we combine these descriptors with domain-specific keypoints, particularly blood vessel bifurcations and crossovers, that are detected using a deep neural network. Our experimental results demonstrate the benefits of the novel multi-positive multi-negative strategy, as it outperforms the widely used triplet loss technique (single-positive and single-negative) as well as the single-positive multi-negative alternative. Additionally, the combination of ConKeD with the domain-specific keypoints produces comparable results to the state-of-the-art methods for retinal image registration, while offering important advantages such as avoiding pre-processing, utilizing fewer training samples, and requiring fewer detected keypoints, among others. Therefore, ConKeD shows a promising potential towards facilitating the development and application of deep learning-based methods for retinal image registration.
Authors: Liwen Hu, Ziluo Ding, Mianzhi Liu, Lei Ma, Tiejun Huang
Abstract: As a neuromorphic sensor with high temporal resolution, spike camera can generate continuous binary spike streams to capture per-pixel light intensity. We can use reconstruction methods to restore scene details in high-speed scenarios. However, due to limited information in spike streams, low-light scenes are difficult to effectively reconstruct. In this paper, we propose a bidirectional recurrent-based reconstruction framework, including a Light-Robust Representation (LR-Rep) and a fusion module, to better handle such extreme conditions. LR-Rep is designed to aggregate temporal information in spike streams, and a fusion module is utilized to extract temporal features. Additionally, we have developed a reconstruction benchmark for high-speed low-light scenes. Light sources in the scenes are carefully aligned to real-world conditions. Experimental results demonstrate the superiority of our method, which also generalizes well to real spike streams. Related codes and proposed datasets will be released after publication.
Authors: Bin Lin, Zhenyu Tang, Yang Ye, Jiaxi Cui, Bin Zhu, Peng Jin, Jinfa Huang, Junwu Zhang, Yatian Pang, Munan Ning, Li Yuan
Abstract: Recent advances demonstrate that scaling Large Vision-Language Models (LVLMs) effectively improves downstream task performances. However, existing scaling methods enable all model parameters to be active for each token in the calculation, which brings massive training and inferring costs. In this work, we propose a simple yet effective training strategy MoE-Tuning for LVLMs. This strategy innovatively addresses the common issue of performance degradation in multi-modal sparsity learning, consequently constructing a sparse model with an outrageous number of parameters but a constant computational cost. Furthermore, we present the MoE-LLaVA, a MoE-based sparse LVLM architecture, which uniquely activates only the top-k experts through routers during deployment, keeping the remaining experts inactive. Extensive experiments show the significant performance of MoE-LLaVA in a variety of visual understanding and object hallucination benchmarks. Remarkably, with only approximately 3B sparsely activated parameters, MoE-LLaVA demonstrates performance comparable to the LLaVA-1.5-7B on various visual understanding datasets and even surpasses the LLaVA-1.5-13B in object hallucination benchmark. Through MoE-LLaVA, we aim to establish a baseline for sparse LVLMs and provide valuable insights for future research in developing more efficient and effective multi-modal learning systems. Code is released at https://github.com/PKU-YuanGroup/MoE-LLaVA.
Authors: Marcos V. Conde, Gregor Geigle, Radu Timofte
Abstract: Image restoration is a fundamental problem that involves recovering a high-quality clean image from its degraded observation. All-In-One image restoration models can effectively restore images from various types and levels of degradation using degradation-specific information as prompts to guide the restoration model. In this work, we present the first approach that uses human-written instructions to guide the image restoration model. Given natural language prompts, our model can recover high-quality images from their degraded counterparts, considering multiple degradation types. Our method, InstructIR, achieves state-of-the-art results on several restoration tasks including image denoising, deraining, deblurring, dehazing, and (low-light) image enhancement. InstructIR improves +1dB over previous all-in-one restoration methods. Moreover, our dataset and results represent a novel benchmark for new research on text-guided image restoration and enhancement. Our code, datasets and models are available at: https://github.com/mv-lab/InstructIR
Authors: Yixuan Wu, Kaiyuan Hu, Danny Z. Chen, Jian Wu
Abstract: With the rapid advance of computer graphics and artificial intelligence technologies, the ways we interact with the world have undergone a transformative shift. Virtual Reality (VR) technology, aided by artificial intelligence (AI), has emerged as a dominant interaction media in multiple application areas, thanks to its advantage of providing users with immersive experiences. Among those applications, medicine is considered one of the most promising areas. In this paper, we present a comprehensive examination of the burgeoning field of AI-enhanced VR applications in medical care and services. By introducing a systematic taxonomy, we meticulously classify the pertinent techniques and applications into three well-defined categories based on different phases of medical diagnosis and treatment: Visualization Enhancement, VR-related Medical Data Processing, and VR-assisted Intervention. This categorization enables a structured exploration of the diverse roles that AI-powered VR plays in the medical domain, providing a framework for a more comprehensive understanding and evaluation of these technologies. To our best knowledge, this is the first systematic survey of AI-powered VR systems in medical settings, laying a foundation for future research in this interdisciplinary domain.
Authors: Hochul Hwang, Sunjae Kwon, Yekyung Kim, Donghyun Kim
Abstract: Safely navigating street intersections is a complex challenge for blind and low-vision individuals, as it requires a nuanced understanding of the surrounding context - a task heavily reliant on visual cues. Traditional methods for assisting in this decision-making process often fall short, lacking the ability to provide a comprehensive scene analysis and safety level. This paper introduces an innovative approach that leverages large multimodal models (LMMs) to interpret complex street crossing scenes, offering a potential advancement over conventional traffic signal recognition techniques. By generating a safety score and scene description in natural language, our method supports safe decision-making for the blind and low-vision individuals. We collected crosswalk intersection data that contains multiview egocentric images captured by a quadruped robot and annotated the images with corresponding safety scores based on our predefined safety score categorization. Grounded on the visual knowledge, extracted from images, and text prompt, we evaluate a large multimodal model for safety score prediction and scene description. Our findings highlight the reasoning and safety score prediction capabilities of a LMM, activated by various prompts, as a pathway to developing a trustworthy system, crucial for applications requiring reliable decision-making support.
Authors: Ufuk Can Bicici, Tuna Han Salih Meral, Lale Akarun
Abstract: Conditional computing processes an input using only part of the neural network's computational units. Learning to execute parts of a deep convolutional network by routing individual samples has several advantages: Reducing the computational burden is an obvious advantage. Furthermore, if similar classes are routed to the same path, that part of the network learns to discriminate between finer differences and better classification accuracies can be attained with fewer parameters. Recently, several papers have exploited this idea to take a particular child of a node in a tree-shaped network or to skip parts of a network. In this work, we follow a Trellis-based approach for generating specific execution paths in a deep convolutional neural network. We have designed routing mechanisms that use differentiable information gain-based cost functions to determine which subset of features in a convolutional layer will be executed. We call our method Conditional Information Gain Trellis (CIGT). We show that our conditional execution mechanism achieves comparable or better model performance compared to unconditional baselines, using only a fraction of the computational resources.
Authors: Zongyu Wu, Hongcheng Gao, Yueze Wang, Xiang Zhang, Suhang Wang
Abstract: Text-to-Image (T2I) models have shown great performance in generating images based on textual prompts. However, these models are vulnerable to unsafe input to generate unsafe content like sexual, harassment and illegal-activity images. Existing studies based on image checker, model fine-tuning and embedding blocking are impractical in real-world applications. Hence, we propose the first universal prompt optimizer for safe T2I (POSI) generation in black-box scenario. We first construct a dataset consisting of toxic-clean prompt pairs by GPT-3.5 Turbo. To guide the optimizer to have the ability of converting toxic prompt to clean prompt while preserving semantic information, we design a novel reward function measuring toxicity and text alignment of generated images and train the optimizer through Proximal Policy Optimization. Experiments show that our approach can effectively reduce the likelihood of various T2I models in generating inappropriate images, with no significant impact on text alignment. It is also flexible to be combined with methods to achieve better performance. Our code is available at https://github.com/wzongyu/POSI.
Authors: Ji Ma, Hongming Dai, Yao Mu, Pengying Wu, Hao Wang, Xiaowei Chi, Yang Fei, Shanghang Zhang, Chang Liu
Abstract: Zero-Shot Object Navigation (ZSON) requires agents to autonomously locate and approach unseen objects in unfamiliar environments and has emerged as a particularly challenging task within the domain of Embodied AI. Existing datasets for developing ZSON algorithms lack consideration of dynamic obstacles, object attribute diversity, and scene texts, thus exhibiting noticeable discrepancies from real-world situations. To address these issues, we propose a Dataset for Open-Vocabulary Zero-Shot Object Navigation in Dynamic Environments (DOZE) that comprises ten high-fidelity 3D scenes with over 18k tasks, aiming to mimic complex, dynamic real-world scenarios. Specifically, DOZE scenes feature multiple moving humanoid obstacles, a wide array of open-vocabulary objects, diverse distinct-attribute objects, and valuable textual hints. Besides, different from existing datasets that only provide collision checking between the agent and static obstacles, we enhance DOZE by integrating capabilities for detecting collisions between the agent and moving obstacles. This novel functionality enables the evaluation of the agents' collision avoidance abilities in dynamic environments. We test four representative ZSON methods on DOZE, revealing substantial room for improvement in existing approaches concerning navigation efficiency, safety, and object recognition accuracy. Our dataset can be found at https://DOZE-Dataset.github.io/.
Authors: Christos Koutlis, Symeon Papadopoulos
Abstract: The recently developed and publicly available synthetic image generation methods and services make it possible to create extremely realistic imagery on demand, raising great risks for the integrity and safety of online information. State-of-the-art Synthetic Image Detection (SID) research has led to strong evidence on the advantages of feature extraction from foundation models. However, such extracted features mostly encapsulate high-level visual semantics instead of fine-grained details, which are more important for the SID task. On the contrary, shallow layers encode low-level visual information. In this work, we leverage the image representations extracted by intermediate Transformer blocks of CLIP's image-encoder via a lightweight network that maps them to a learnable forgery-aware vector space capable of generalizing exceptionally well. We also employ a trainable module to incorporate the importance of each Transformer block to the final prediction. Our method is compared against the state-of-the-art by evaluating it on 20 test datasets and exhibits an average +10.6% absolute performance improvement. Notably, the best performing models require just a single epoch for training (~8 minutes). Code available at https://github.com/mever-team/rine.
Authors: Xiangxiang Chu, Jianlin Su, Bo Zhang, Chunhua Shen
Abstract: Large language models are built on top of a transformer-based architecture to process textual inputs. For example, the LLaMA stands out among many open-source implementations. Can the same transformer be used to process 2D images? In this paper, we answer this question by unveiling a LLaMA-like vision transformer in plain and pyramid forms, termed VisionLLaMA, which is tailored for this purpose. VisionLLaMA is a unified and generic modelling framework for solving most vision tasks. We extensively evaluate its effectiveness using typical pre-training paradigms in a good portion of downstream tasks of image perception and especially image generation. In many cases, VisionLLaMA have exhibited substantial gains over the previous state-of-the-art vision transformers. We believe that VisionLLaMA can serve as a strong new baseline model for vision generation and understanding. Our code is released at https://github.com/Meituan-AutoML/VisionLLaMA.
Authors: Quan Chen, Tingyu Wang, Zihao Yang, Haoran Li, Rongfeng Lu, Yaoqi Sun, Bolun Zheng, Chenggang Yan
Abstract: Cross-view geo-localization aims to match images of the same target from different platforms, e.g., drone and satellite. It is a challenging task due to the changing appearance of targets and environmental content from different views. Most methods focus on obtaining more comprehensive information through feature map segmentation, while inevitably destroying the image structure, and are sensitive to the shifting and scale of the target in the query. To address the above issues, we introduce simple yet effective part-based representation learning, shifting-dense partition learning (SDPL). We propose a dense partition strategy (DPS), dividing the image into multiple parts to explore contextual information while explicitly maintaining the global structure. To handle scenarios with non-centered targets, we further propose the shifting-fusion strategy, which generates multiple sets of parts in parallel based on various segmentation centers, and then adaptively fuses all features to integrate their anti-offset ability. Extensive experiments show that SDPL is robust to position shifting, and performs com-petitively on two prevailing benchmarks, University-1652 and SUES-200. In addition, SDPL shows satisfactory compatibility with a variety of backbone networks (e.g., ResNet and Swin). https://github.com/C-water/SDPL release.
Authors: Jiawei Chen, Yue Jiang, Dingkang Yang, Mingcheng Li, Jinjie Wei, Ziyun Qian, Lihua Zhang
Abstract: While Large Language Models (LLMs) excel in world knowledge understanding, adapting them to specific subfields requires precise adjustments. Due to the model's vast scale, traditional global fine-tuning methods for large models can be computationally expensive and impact generalization. To address this challenge, a range of innovative Parameters-Efficient Fine-Tuning (PEFT) methods have emerged and achieved remarkable success in both LLMs and Large Vision-Language Models (LVLMs). In the medical domain, fine-tuning a medical Vision-Language Pretrained (VLP) model is essential for adapting it to specific tasks. Can the fine-tuning methods for large models be transferred to the medical field to enhance transfer learning efficiency? In this paper, we delve into the fine-tuning methods of LLMs and conduct extensive experiments to investigate the impact of fine-tuning methods for large models on the existing multimodal model in the medical domain from the training data level and the model structure level. We show the different impacts of fine-tuning methods for large models on medical VLMs and develop the most efficient ways to fine-tune medical VLP models. We hope this research can guide medical domain researchers in optimizing VLMs' training costs, fostering the broader application of VLMs in healthcare fields. The code and dataset have been released at https://github.com/TIMMY-CHAN/MILE.
Authors: Hayeon O, Chanuk Yang, Kunsoo Huh
Abstract: In autonomous driving, 3D object detection provides more precise information for downstream tasks, including path planning and motion estimation, compared to 2D object detection. In this paper, we propose SeSame: a method aimed at enhancing semantic information in existing LiDAR-only based 3D object detection. This addresses the limitation of existing 3D detectors, which primarily focus on object presence and classification, thus lacking in capturing relationships between elemental units that constitute the data, akin to semantic segmentation. Experiments demonstrate the effectiveness of our method with performance improvements on the KITTI object detection benchmark. Our code is available at https://github.com/HAMA-DL-dev/SeSame
Authors: Hao Shi, Song Wang, Jiaming Zhang, Xiaoting Yin, Zhongdao Wang, Guangming Wang, Jianke Zhu, Kailun Yang, Kaiwei Wang
Abstract: Vision-based occupancy prediction, also known as 3D Semantic Scene Completion (SSC), presents a significant challenge in computer vision. Previous methods, confined to onboard processing, struggle with simultaneous geometric and semantic estimation, continuity across varying viewpoints, and single-view occlusion. Our paper introduces OccFiner, a novel offboard framework designed to enhance the accuracy of vision-based occupancy predictions. OccFiner operates in two hybrid phases: 1) a multi-to-multi local propagation network that implicitly aligns and processes multiple local frames for correcting onboard model errors and consistently enhancing occupancy accuracy across all distances. 2) the region-centric global propagation, focuses on refining labels using explicit multi-view geometry and integrating sensor bias, especially to increase the accuracy of distant occupied voxels. Extensive experiments demonstrate that OccFiner improves both geometric and semantic accuracy across various types of coarse occupancy, setting a new state-of-the-art performance on the SemanticKITTI dataset. Notably, OccFiner elevates vision-based SSC models to a level even surpassing that of LiDAR-based onboard SSC models. Furthermore, OccFiner is the first to achieve automatic annotation of SSC in a purely vision-based approach. Quantitative experiments prove that OccFiner successfully facilitates occupancy data loop-closure in autonomous driving. Additionally, we quantitatively and qualitatively validate the superiority of the offboard approach on city-level SSC static maps. The source code will be made publicly available at https://github.com/MasterHow/OccFiner.
Authors: Jiyuan Fu, Zhaoyu Chen, Kaixun Jiang, Haijing Guo, Jiafeng Wang, Shuyong Gao, Wenqiang Zhang
Abstract: Despite the substantial advancements in Vision-Language Pre-training (VLP) models, their susceptibility to adversarial attacks poses a significant challenge. Existing work rarely studies the transferability of attacks on VLP models, resulting in a substantial performance gap from white-box attacks. We observe that prior work overlooks the interaction mechanisms between modalities, which plays a crucial role in understanding the intricacies of VLP models. In response, we propose a novel attack, called Collaborative Multimodal Interaction Attack (CMI-Attack), leveraging modality interaction through embedding guidance and interaction enhancement. Specifically, attacking text at the embedding level while preserving semantics, as well as utilizing interaction image gradients to enhance constraints on perturbations of texts and images. Significantly, in the image-text retrieval task on Flickr30K dataset, CMI-Attack raises the transfer success rates from ALBEF to TCL, $\text{CLIP}_{\text{ViT}}$ and $\text{CLIP}_{\text{CNN}}$ by 8.11%-16.75% over state-of-the-art methods. Moreover, CMI-Attack also demonstrates superior performance in cross-task generalization scenarios. Our work addresses the underexplored realm of transfer attacks on VLP models, shedding light on the importance of modality interaction for enhanced adversarial robustness.
Authors: Zixin Zhu, Xuelu Feng, Dongdong Chen, Junsong Yuan, Chunming Qiao, Gang Hua
Abstract: In this paper, we explore the visual representations produced from a pre-trained text-to-video (T2V) diffusion model for video understanding tasks. We hypothesize that the latent representation learned from a pretrained generative T2V model encapsulates rich semantics and coherent temporal correspondences, thereby naturally facilitating video understanding. Our hypothesis is validated through the classic referring video object segmentation (R-VOS) task. We introduce a novel framework, termed "VD-IT", tailored with dedicatedly designed components built upon a fixed pretrained T2V model. Specifically, VD-IT uses textual information as a conditional input, ensuring semantic consistency across time for precise temporal instance matching. It further incorporates image tokens as supplementary textual inputs, enriching the feature set to generate detailed and nuanced masks. Besides, instead of using the standard Gaussian noise, we propose to predict the video-specific noise with an extra noise prediction module, which can help preserve the feature fidelity and elevates segmentation quality. Through extensive experiments, we surprisingly observe that fixed generative T2V diffusion models, unlike commonly used video backbones (e.g., Video Swin Transformer) pretrained with discriminative image/video pre-tasks, exhibit better potential to maintain semantic alignment and temporal consistency. On existing standard benchmarks, our VD-IT achieves highly competitive results, surpassing many existing state-of-the-art methods. The code is available at https://github.com/buxiangzhiren/VD-IT.
Authors: Hao Shao, Shengju Qian, Han Xiao, Guanglu Song, Zhuofan Zong, Letian Wang, Yu Liu, Hongsheng Li
Abstract: Multi-Modal Large Language Models (MLLMs) have demonstrated impressive performance in various VQA tasks. However, they often lack interpretability and struggle with complex visual inputs, especially when the resolution of the input image is high or when the interested region that could provide key information for answering the question is small. To address these challenges, we collect and introduce the large-scale Visual CoT dataset comprising 438k question-answer pairs, annotated with intermediate bounding boxes highlighting key regions essential for answering the questions. Additionally, about 98k pairs of them are annotated with detailed reasoning steps. Importantly, we propose a multi-turn processing pipeline that dynamically focuses on visual inputs and provides interpretable thoughts. We also introduce the related benchmark to evaluate the MLLMs in scenarios requiring specific local region identification. Extensive experiments demonstrate the effectiveness of our framework and shed light on better inference strategies. The Visual CoT dataset, benchmark, and pre-trained models are released to foster further research in this direction.
Authors: Yifan Yan, Ruomin He, Zhenghua Liu
Abstract: We introduce MUTE-SLAM, a real-time neural RGB-D SLAM system employing multiple tri-plane hash-encodings for efficient scene representation. MUTE-SLAM effectively tracks camera positions and incrementally builds a scalable multi-map representation for both small and large indoor environments. As previous methods often require pre-defined scene boundaries, MUTE-SLAM dynamically allocates sub-maps for newly observed local regions, enabling constraint-free mapping without prior scene information. Unlike traditional grid-based methods, we use three orthogonal axis-aligned planes for hash-encoding scene properties, significantly reducing hash collisions and the number of trainable parameters. This hybrid approach not only ensures real-time performance but also enhances the fidelity of surface reconstruction. Furthermore, our optimization strategy concurrently optimizes all sub-maps intersecting with the current camera frustum, ensuring global consistency. Extensive testing on both real-world and synthetic datasets has shown that MUTE-SLAM delivers state-of-the-art surface reconstruction quality and competitive tracking performance across diverse indoor settings. The code will be made public upon acceptance of the paper.
Authors: Yang Miao, Francis Engelmann, Olga Vysotska, Federico Tombari, Marc Pollefeys, D\'aniel B\'ela Bar\'ath
Abstract: We introduce a novel problem, i.e., the localization of an input image within a multi-modal reference map represented by a database of 3D scene graphs. These graphs comprise multiple modalities, including object-level point clouds, images, attributes, and relationships between objects, offering a lightweight and efficient alternative to conventional methods that rely on extensive image databases. Given the available modalities, the proposed method SceneGraphLoc learns a fixed-sized embedding for each node (i.e., representing an object instance) in the scene graph, enabling effective matching with the objects visible in the input query image. This strategy significantly outperforms other cross-modal methods, even without incorporating images into the map embeddings. When images are leveraged, SceneGraphLoc achieves performance close to that of state-of-the-art techniques depending on large image databases, while requiring three orders-of-magnitude less storage and operating orders-of-magnitude faster. The code will be made public.
Authors: Tanvir Mahmud, Yapeng Tian, Diana Marculescu
Abstract: Visual sound source localization poses a significant challenge in identifying the semantic region of each sounding source within a video. Existing self-supervised and weakly supervised source localization methods struggle to accurately distinguish the semantic regions of each sounding object, particularly in multi-source mixtures. These methods often rely on audio-visual correspondence as guidance, which can lead to substantial performance drops in complex multi-source localization scenarios. The lack of access to individual source sounds in multi-source mixtures during training exacerbates the difficulty of learning effective audio-visual correspondence for localization. To address this limitation, in this paper, we propose incorporating the text modality as an intermediate feature guide using tri-modal joint embedding models (e.g., AudioCLIP) to disentangle the semantic audio-visual source correspondence in multi-source mixtures. Our framework, dubbed T-VSL, begins by predicting the class of sounding entities in mixtures. Subsequently, the textual representation of each sounding source is employed as guidance to disentangle fine-grained audio-visual source correspondence from multi-source mixtures, leveraging the tri-modal AudioCLIP embedding. This approach enables our framework to handle a flexible number of sources and exhibits promising zero-shot transferability to unseen classes during test time. Extensive experiments conducted on the MUSIC, VGGSound, and VGGSound-Instruments datasets demonstrate significant performance improvements over state-of-the-art methods. Code is released at https://github.com/enyac-group/T-VSL/tree/main
Authors: Magamed Taimeskhanov, Ronan Sicre, Damien Garreau
Abstract: CAM-based methods are widely-used post-hoc interpretability method that produce a saliency map to explain the decision of an image classification model. The saliency map highlights the important areas of the image relevant to the prediction. In this paper, we show that most of these methods can incorrectly attribute an important score to parts of the image that the model cannot see. We show that this phenomenon occurs both theoretically and experimentally. On the theory side, we analyze the behavior of GradCAM on a simple masked CNN model at initialization. Experimentally, we train a VGG-like model constrained to not use the lower part of the image and nevertheless observe positive scores in the unseen part of the image. This behavior is evaluated quantitatively on two new datasets. We believe that this is problematic, potentially leading to mis-interpretation of the model's behavior.
Authors: Xiaoshuang Huang, Hongxiang Li, Meng Cao, Long Chen, Chenyu You, Dong An
Abstract: Recent developments underscore the potential of textual information in enhancing learning models for a deeper understanding of medical visual semantics. However, language-guided medical image segmentation still faces a challenging issue. Previous works employ implicit and ambiguous architectures to embed textual information. This leads to segmentation results that are inconsistent with the semantics represented by the language, sometimes even diverging significantly. To this end, we propose a novel cross-modal conditioned Reconstruction for Language-guided Medical Image Segmentation (RecLMIS) to explicitly capture cross-modal interactions, which assumes that well-aligned medical visual features and medical notes can effectively reconstruct each other. We introduce conditioned interaction to adaptively predict patches and words of interest. Subsequently, they are utilized as conditioning factors for mutual reconstruction to align with regions described in the medical notes. Extensive experiments demonstrate the superiority of our RecLMIS, surpassing LViT by 3.74% mIoU on the publicly available MosMedData+ dataset and achieving an average increase of 1.89% mIoU for cross-domain tests on our QATA-CoV19 dataset. Simultaneously, we achieve a relative reduction of 20.2% in parameter count and a 55.5% decrease in computational load. The code will be available at https://github.com/ShashankHuang/RecLMIS.
Authors: Seunghoi Kim, Chen Jin, Tom Diethe, Matteo Figini, Henry F. J. Tregidgo, Asher Mullokandov, Philip Teare, Daniel C. Alexander
Abstract: Recent developments in diffusion models have advanced conditioned image generation, yet they struggle with reconstructing out-of-distribution (OOD) images, such as unseen tumors in medical images, causing "image hallucination" and risking misdiagnosis. We hypothesize such hallucinations result from local OOD regions in the conditional images. We verify that partitioning the OOD region and conducting separate image generations alleviates hallucinations in several applications. From this, we propose a training-free diffusion framework that reduces hallucination with multiple Local Diffusion processes. Our approach involves OOD estimation followed by two modules: a "branching" module generates locally both within and outside OOD regions, and a "fusion" module integrates these predictions into one. Our evaluation shows our method mitigates hallucination over baseline models quantitatively and qualitatively, reducing misdiagnosis by 40% and 25% in the real-world medical and natural image datasets, respectively. It also demonstrates compatibility with various pre-trained diffusion models.
Authors: Anas Gouda, Max Schwarz, Christopher Reining, Sven Behnke, Alice Kirchheim
Abstract: Foundation models are a strong trend in deep learning and computer vision. These models serve as a base for applications as they require minor or no further fine-tuning by developers to integrate into their applications. Foundation models for zero-shot object segmentation such as Segment Anything (SAM) output segmentation masks from images without any further object information. When they are followed in a pipeline by an object identification model, they can perform object detection without training. Here, we focus on training such an object identification model. A crucial practical aspect for an object identification model is to be flexible in input size. As object identification is an image retrieval problem, a suitable method should handle multi-query multi-gallery situations without constraining the number of input images (e.g. by having fixed-size aggregation layers). The key solution to train such a model is the centroid triplet loss (CTL), which aggregates image features to their centroids. CTL yields high accuracy, avoids misleading training signals and keeps the model input size flexible. In our experiments, we establish a new state of the art on the ArmBench object identification task, which shows general applicability of our model. We furthermore demonstrate an integrated unseen object detection pipeline on the challenging HOPE dataset, which requires fine-grained detection. There, our pipeline matches and surpasses related methods which have been trained on dataset-specific data.
Authors: Robin Weiler, Matthias Brucklacher, Cyriel M. A. Pennartz, Sander M. Boht\'e
Abstract: To make sense of their surroundings, intelligent systems must transform complex sensory inputs to structured codes that are reduced to task-relevant information such as object category. Biological agents achieve this in a largely autonomous manner, presumably via self-supervised learning. Whereas previous attempts to model the underlying mechanisms were largely discriminative in nature, there is ample evidence that the brain employs a generative model of the world. Here, we propose that eye movements, in combination with the focused nature of primate vision, constitute a generative, self-supervised task of predicting and revealing visual information. We construct a proof-of-principle model starting from the framework of masked image modeling (MIM), a common approach in deep representation learning. To do so, we analyze how core components of MIM such as masking technique and data augmentation influence the formation of category-specific representations. This allows us not only to better understand the principles behind MIM, but to then reassemble a MIM more in line with the focused nature of biological perception. We find that MIM disentangles neurons in latent space without explicit regularization, a property that has been suggested to structure visual representations in primates. Together with previous findings of invariance learning, this highlights an interesting connection of MIM to latent regularization approaches for self-supervised learning. The source code is available under https://github.com/RobinWeiler/FocusMIM
Authors: Akshatha Mohan, Joshua Peeples
Abstract: Pooling layers (e.g., max and average) may overlook important information encoded in the spatial arrangement of pixel intensity and/or feature values. We propose a novel lacunarity pooling layer that aims to capture the spatial heterogeneity of the feature maps by evaluating the variability within local windows. The layer operates at multiple scales, allowing the network to adaptively learn hierarchical features. The lacunarity pooling layer can be seamlessly integrated into any artificial neural network architecture. Experimental results demonstrate the layer's effectiveness in capturing intricate spatial patterns, leading to improved feature extraction capabilities. The proposed approach holds promise in various domains, especially in agricultural image analysis tasks. This work contributes to the evolving landscape of artificial neural network architectures by introducing a novel pooling layer that enriches the representation of spatial features. Our code is publicly available.
Authors: You Wu, Yuelong Wang, Yaxin Liao, Fuliang Wu, Hengzhou Ye, Shuiwang Li
Abstract: Tracking transforming objects holds significant importance in various fields due to the dynamic nature of many real-world scenarios. By enabling systems accurately represent transforming objects over time, tracking transforming objects facilitates advancements in areas such as autonomous systems, human-computer interaction, and security applications. Moreover, understanding the behavior of transforming objects provides valuable insights into complex interactions or processes, contributing to the development of intelligent systems capable of robust and adaptive perception in dynamic environments. However, current research in the field mainly focuses on tracking generic objects. In this study, we bridge this gap by collecting a novel dedicated Dataset for Tracking Transforming Objects, called DTTO, which contains 100 sequences, amounting to approximately 9.3K frames. We provide carefully hand-annotated bounding boxes for each frame within these sequences, making DTTO the pioneering benchmark dedicated to tracking transforming objects. We thoroughly evaluate 20 state-of-the-art trackers on the benchmark, aiming to comprehend the performance of existing methods and provide a comparison for future research on DTTO. With the release of DTTO, our goal is to facilitate further research and applications related to tracking transforming objects.
Authors: Rui Xu, Shu Yang, Yihui Wang, Yu Cai, Bo Du, Hao Chen
Abstract: Mamba, a recent selective structured state space model, excels in long sequence modeling, which is vital in the large model era. Long sequence modeling poses significant challenges, including capturing long-range dependencies within the data and handling the computational demands caused by their extensive length. Mamba addresses these challenges by overcoming the local perception limitations of convolutional neural networks and the quadratic computational complexity of Transformers. Given its advantages over these mainstream foundation architectures, Mamba exhibits great potential to be a visual foundation architecture. Since January 2024, Mamba has been actively applied to diverse computer vision tasks, yielding numerous contributions. To help keep pace with the rapid advancements, this paper reviews visual Mamba approaches, analyzing over 200 papers. This paper begins by delineating the formulation of the original Mamba model. Subsequently, it delves into representative backbone networks, and applications categorized using different modalities, including image, video, point cloud, and multi-modal. Particularly, we identify scanning techniques as critical for adapting Mamba to vision tasks, and decouple these scanning techniques to clarify their functionality and enhance their flexibility across various applications. Finally, we discuss the challenges and future directions, providing insights into new outlooks in this fast evolving area. A comprehensive list of visual Mamba models reviewed in this work is available at https://github.com/Ruixxxx/Awesome-Vision-Mamba-Models.
URLs: https://github.com/Ruixxxx/Awesome-Vision-Mamba-Models.
Authors: Teng Zhou, Yongchuan Tang
Abstract: Diffusion models have emerged as effective tools for generating diverse and high-quality content. However, their capability in high-resolution image generation, particularly for panoramic images, still faces challenges such as visible seams and incoherent transitions. In this paper, we propose TwinDiffusion, an optimized framework designed to address these challenges through two key innovations: the Crop Fusion for quality enhancement and the Cross Sampling for efficiency optimization. We introduce a training-free optimizing stage to refine the similarity of adjacent image areas, as well as an interleaving sampling strategy to yield dynamic patches during the cropping process. A comprehensive evaluation is conducted to compare TwinDiffusion with the prior works, considering factors including coherence, fidelity, compatibility, and efficiency. The results demonstrate the superior performance of our approach in generating seamless and coherent panoramas, setting a new standard in quality and efficiency for panoramic image generation.
Authors: Yanan Zhang, Jinqing Zhang, Zengran Wang, Junhao Xu, Di Huang
Abstract: In recent years, autonomous driving has garnered escalating attention for its potential to relieve drivers' burdens and improve driving safety. Vision-based 3D occupancy prediction, which predicts the spatial occupancy status and semantics of 3D voxel grids around the autonomous vehicle from image inputs, is an emerging perception task suitable for cost-effective perception system of autonomous driving. Although numerous studies have demonstrated the greater advantages of 3D occupancy prediction over object-centric perception tasks, there is still a lack of a dedicated review focusing on this rapidly developing field. In this paper, we first introduce the background of vision-based 3D occupancy prediction and discuss the challenges in this task. Secondly, we conduct a comprehensive survey of the progress in vision-based 3D occupancy prediction from three aspects: feature enhancement, deployment friendliness and label efficiency, and provide an in-depth analysis of the potentials and challenges of each category of methods. Finally, we present a summary of prevailing research trends and propose some inspiring future outlooks. To provide a valuable reference for researchers, a regularly updated collection of related papers, datasets, and codes is organized at https://github.com/zya3d/Awesome-3D-Occupancy-Prediction.
URLs: https://github.com/zya3d/Awesome-3D-Occupancy-Prediction.
Authors: Faegheh Sardari, Armin Mustafa, Philip J. B. Jackson, Adrian Hilton
Abstract: Weakly supervised audio-visual video parsing (AVVP) methods aim to detect audible-only, visible-only, and audible-visible events using only video-level labels. Existing approaches tackle this by leveraging unimodal and cross-modal contexts. However, we argue that while cross-modal learning is beneficial for detecting audible-visible events, in the weakly supervised scenario, it negatively impacts unaligned audible or visible events by introducing irrelevant modality information. In this paper, we propose CoLeaF, a novel learning framework that optimizes the integration of cross-modal context in the embedding space such that the network explicitly learns to combine cross-modal information for audible-visible events while filtering them out for unaligned events. Additionally, as videos often involve complex class relationships, modelling them improves performance. However, this introduces extra computational costs into the network. Our framework is designed to leverage cross-class relationships during training without incurring additional computations at inference. Furthermore, we propose new metrics to better evaluate a method's capabilities in performing AVVP. Our extensive experiments demonstrate that CoLeaF significantly improves the state-of-the-art results by an average of 1.9% and 2.4% F-score on the LLP and UnAV-100 datasets, respectively.
Authors: Tianqi Liu, Guangcong Wang, Shoukang Hu, Liao Shen, Xinyi Ye, Yuhang Zang, Zhiguo Cao, Wei Li, Ziwei Liu
Abstract: We present MVSGaussian, a new generalizable 3D Gaussian representation approach derived from Multi-View Stereo (MVS) that can efficiently reconstruct unseen scenes. Specifically, 1) we leverage MVS to encode geometry-aware Gaussian representations and decode them into Gaussian parameters. 2) To further enhance performance, we propose a hybrid Gaussian rendering that integrates an efficient volume rendering design for novel view synthesis. 3) To support fast fine-tuning for specific scenes, we introduce a multi-view geometric consistent aggregation strategy to effectively aggregate the point clouds generated by the generalizable model, serving as the initialization for per-scene optimization. Compared with previous generalizable NeRF-based methods, which typically require minutes of fine-tuning and seconds of rendering per image, MVSGaussian achieves real-time rendering with better synthesis quality for each scene. Compared with the vanilla 3D-GS, MVSGaussian achieves better view synthesis with less training computational cost. Extensive experiments on DTU, Real Forward-facing, NeRF Synthetic, and Tanks and Temples datasets validate that MVSGaussian attains state-of-the-art performance with convincing generalizability, real-time rendering speed, and fast per-scene optimization.
Authors: Haonan Zhang, Pengpeng Zeng, Lianli Gao, Jingkuan Song, Yihang Duan, Xinyu Lyu, Hengtao Shen
Abstract: Adapting large-scale image-text pre-training models, e.g., CLIP, to the video domain represents the current state-of-the-art for text-video retrieval. The primary approaches involve transferring text-video pairs to a common embedding space and leveraging cross-modal interactions on specific entities for semantic alignment. Though effective, these paradigms entail prohibitive computational costs, leading to inefficient retrieval. To address this, we propose a simple yet effective method, Global-Local Semantic Consistent Learning (GLSCL), which capitalizes on latent shared semantics across modalities for text-video retrieval. Specifically, we introduce a parameter-free global interaction module to explore coarse-grained alignment. Then, we devise a shared local interaction module that employs several learnable queries to capture latent semantic concepts for learning fine-grained alignment. Furthermore, an Inter-Consistency Loss (ICL) is devised to accomplish the concept alignment between the visual query and corresponding textual query, and an Intra-Diversity Loss (IDL) is developed to repulse the distribution within visual (textual) queries to generate more discriminative concepts. Extensive experiments on five widely used benchmarks (i.e., MSR-VTT, MSVD, DiDeMo, LSMDC, and ActivityNet) substantiate the superior effectiveness and efficiency of the proposed method. Remarkably, our method achieves comparable performance with SOTA as well as being nearly 220 times faster in terms of computational cost. Code is available at: https://github.com/zchoi/GLSCL.
Authors: Neehar Kondapaneni, Markus Marks, Oisin Mac Aodha, Pietro Perona
Abstract: We introduce Discovering Conceptual Network Explanations (DCNE), a new approach for generating human-comprehensible visual explanations to enhance the interpretability of deep neural image classifiers. Our method automatically finds visual explanations that are critical for discriminating between classes. This is achieved by simultaneously optimizing three criteria: the explanations should be few, diverse, and human-interpretable. Our approach builds on the recently introduced Concept Relevance Propagation (CRP) explainability method. While CRP is effective at describing individual neuronal activations, it generates too many concepts, which impacts human comprehension. Instead, DCNE selects the few most important explanations. We introduce a new evaluation dataset centered on the challenging task of classifying birds, enabling us to compare the alignment of DCNE's explanations to those of human expert-defined ones. Compared to existing eXplainable Artificial Intelligence (XAI) methods, DCNE has a desirable trade-off between conciseness and completeness when summarizing network explanations. It produces 1/30 of CRP's explanations while only resulting in a slight reduction in explanation quality. DCNE represents a step forward in making neural network decisions accessible and interpretable to humans, providing a valuable tool for both researchers and practitioners in XAI and model alignment.
Authors: Zhoujie Fu, Jiacheng Wei, Wenhao Shen, Chaoyue Song, Xiaofeng Yang, Fayao Liu, Xulei Yang, Guosheng Lin
Abstract: In this work, we introduce a novel approach for creating controllable dynamics in 3D-generated Gaussians using casually captured reference videos. Our method transfers the motion of objects from reference videos to a variety of generated 3D Gaussians across different categories, ensuring precise and customizable motion transfer. We achieve this by employing blend skinning-based non-parametric shape reconstruction to extract the shape and motion of reference objects. This process involves segmenting the reference objects into motion-related parts based on skinning weights and establishing shape correspondences with generated target shapes. To address shape and temporal inconsistencies prevalent in existing methods, we integrate physical simulation, driving the target shapes with matched motion. This integration is optimized through a displacement loss to ensure reliable and genuine dynamics. Our approach supports diverse reference inputs, including humans, quadrupeds, and articulated objects, and can generate dynamics of arbitrary length, providing enhanced fidelity and applicability. Unlike methods heavily reliant on diffusion video generation models, our technique offers specific and high-quality motion transfer, maintaining both shape integrity and temporal consistency.
Authors: Weitai Kang, Mengxue Qu, Jyoti Kini, Yunchao Wei, Mubarak Shah, Yan Yan
Abstract: In real-life scenarios, humans seek out objects in the 3D world to fulfill their daily needs or intentions. This inspires us to introduce 3D intention grounding, a new task in 3D object detection employing RGB-D, based on human intention, such as "I want something to support my back". Closely related, 3D visual grounding focuses on understanding human reference. To achieve detection based on human intention, it relies on humans to observe the scene, reason out the target that aligns with their intention ("pillow" in this case), and finally provide a reference to the AI system, such as "A pillow on the couch". Instead, 3D intention grounding challenges AI agents to automatically observe, reason and detect the desired target solely based on human intention. To tackle this challenge, we introduce the new Intent3D dataset, consisting of 44,990 intention texts associated with 209 fine-grained classes from 1,042 scenes of the ScanNet dataset. We also establish several baselines based on different language-based 3D object detection models on our benchmark. Finally, we propose IntentNet, our unique approach, designed to tackle this intention-based detection problem. It focuses on three key aspects: intention understanding, reasoning to identify object candidates, and cascaded adaptive learning that leverages the intrinsic priority logic of different losses for multiple objective optimization.
Authors: Hongyuan Dong, Jiawen Li, Bohong Wu, Jiacong Wang, Yuan Zhang, Haoyuan Guo
Abstract: Image captioning has long been regarded as a fundamental task in visual understanding. Recently, however, few large vision-language model (LVLM) research discusses model's image captioning performance because of the outdated short-caption benchmarks and unreliable evaluation metrics. In this work, we propose to benchmark detail image caption task by curating high-quality evaluation datasets annotated by human experts, GPT-4V and Gemini-1.5-Pro. We also design a more reliable caption evaluation metric called CAPTURE (CAPtion evaluation by exTracting and coUpling coRE information). CAPTURE extracts visual elements, e.g., objects, attributes and relations from captions, and then matches these elements through three stages, achieving the highest consistency with expert judgements over other rule-based or model-based caption metrics. The proposed benchmark and metric provide reliable evaluation for LVLM's detailed image captioning ability. Guided by this evaluation, we further explore to unleash LVLM's detail caption capabilities by synthesizing high-quality data through a five-stage data construction pipeline. Our pipeline only uses a given LVLM itself and other open-source tools, without any human or GPT-4V annotation in the loop. Experiments show that the proposed data construction strategy significantly improves model-generated detail caption data quality for LVLMs with leading performance, and the data quality can be further improved in a self-looping paradigm. All code and dataset will be publicly available at https://github.com/foundation-multimodal-models/CAPTURE.
URLs: https://github.com/foundation-multimodal-models/CAPTURE.
Authors: Qilong Zhangli, Jindong Jiang, Di Liu, Licheng Yu, Xiaoliang Dai, Ankit Ramchandani, Guan Pang, Dimitris N. Metaxas, Praveen Krishnan
Abstract: While diffusion models have significantly advanced the quality of image generation, their capability to accurately and coherently render text within these images remains a substantial challenge. Conventional diffusion-based methods for scene text generation are typically limited by their reliance on an intermediate layout output. This dependency often results in a constrained diversity of text styles and fonts, an inherent limitation stemming from the deterministic nature of the layout generation phase. To address these challenges, this paper introduces SceneTextGen, a novel diffusion-based model specifically designed to circumvent the need for a predefined layout stage. By doing so, SceneTextGen facilitates a more natural and varied representation of text. The novelty of SceneTextGen lies in its integration of three key components: a character-level encoder for capturing detailed typographic properties, coupled with a character-level instance segmentation model and a word-level spotting model to address the issues of unwanted text generation and minor character inaccuracies. We validate the performance of our method by demonstrating improved character recognition rates on generated images across different public visual text datasets in comparison to both standard diffusion based methods and text specific methods.
Authors: Hoang H. Le, Duy M. H. Nguyen, Omair Shahzad Bhatti, Laszlo Kopacsi, Thinh P. Ngo, Binh T. Nguyen, Michael Barz, Daniel Sonntag
Abstract: Comprehending how humans process visual information in dynamic settings is crucial for psychology and designing user-centered interactions. While mobile eye-tracking systems combining egocentric video and gaze signals can offer valuable insights, manual analysis of these recordings is time-intensive. In this work, we present a novel human-centered learning algorithm designed for automated object recognition within mobile eye-tracking settings. Our approach seamlessly integrates an object detector with a spatial relation-aware inductive message-passing network (I-MPN), harnessing node profile information and capturing object correlations. Such mechanisms enable us to learn embedding functions capable of generalizing to new object angle views, facilitating rapid adaptation and efficient reasoning in dynamic contexts as users navigate their environment. Through experiments conducted on three distinct video sequences, our interactive-based method showcases significant performance improvements over fixed training/testing algorithms, even when trained on considerably smaller annotated samples collected through user feedback. Furthermore, we demonstrate exceptional efficiency in data annotation processes and surpass prior interactive methods that use complete object detectors, combine detectors with convolutional networks, or employ interactive video segmentation.
Authors: Fangliangzi Meng, Hongrun Zhang, Ruodan Yan, Guohui Chuai, Chao Li, Qi Liu
Abstract: The characterization of Tumor MicroEnvironment (TME) is challenging due to its complexity and heterogeneity. Relatively consistent TME characteristics embedded within highly specific tissue features, render them difficult to predict. The capability to accurately classify TME subtypes is of critical significance for clinical tumor diagnosis and precision medicine. Based on the observation that tumors with different origins share similar microenvironment patterns, we propose PathoTME, a genomics-guided Siamese representation learning framework employing Whole Slide Image (WSI) for pan-cancer TME subtypes prediction. Specifically, we utilize Siamese network to leverage genomic information as a regularization factor to assist WSI embeddings learning during the training phase. Additionally, we employ Domain Adversarial Neural Network (DANN) to mitigate the impact of tissue type variations. To eliminate domain bias, a dynamic WSI prompt is designed to further unleash the model's capabilities. Our model achieves better performance than other state-of-the-art methods across 23 cancer types on TCGA dataset. Our code is available at https://github.com/Mengflz/PathoTME.
Authors: Junwei Luo, Zhen Pang, Yongjun Zhang, Tingzhu Wang, Linlin Wang, Bo Dang, Jiangwei Lao, Jian Wang, Jingdong Chen, Yihua Tan, Yansheng Li
Abstract: Remote Sensing Large Multi-Modal Models (RSLMMs) are developing rapidly and showcase significant capabilities in remote sensing imagery (RSI) comprehension. However, due to the limitations of existing datasets, RSLMMs have shortcomings in understanding the rich semantic relations among objects in complex remote sensing scenes. To unlock RSLMMs' complex comprehension ability, we propose a large-scale instruction tuning dataset FIT-RS, containing 1,800,851 instruction samples. FIT-RS covers common interpretation tasks and innovatively introduces several complex comprehension tasks of escalating difficulty, ranging from relation reasoning to image-level scene graph generation. Based on FIT-RS, we build the FIT-RSFG benchmark. Furthermore, we establish a new benchmark to evaluate the fine-grained relation comprehension capabilities of LMMs, named FIT-RSRC. Based on combined instruction data, we propose SkySenseGPT, which achieves outstanding performance on both public datasets and FIT-RSFG, surpassing existing RSLMMs. We hope the FIT-RS dataset can enhance the relation comprehension capability of RSLMMs and provide a large-scale fine-grained data source for the remote sensing community. The dataset will be available at https://github.com/Luo-Z13/SkySenseGPT
Authors: Rishab Parthasarathy, Zachary Ankner, Aaron Gokaslan
Abstract: A recent frontier in computer vision has been the task of 3D video generation, which consists of generating a time-varying 3D representation of a scene. To generate dynamic 3D scenes, current methods explicitly model 3D temporal dynamics by jointly optimizing for consistency across both time and views of the scene. In this paper, we instead investigate whether it is necessary to explicitly enforce multiview consistency over time, as current approaches do, or if it is sufficient for a model to generate 3D representations of each timestep independently. We hence propose a model, Vid3D, that leverages 2D video diffusion to generate 3D videos by first generating a 2D "seed" of the video's temporal dynamics and then independently generating a 3D representation for each timestep in the seed video. We evaluate Vid3D against two state-of-the-art 3D video generation methods and find that Vid3D is achieves comparable results despite not explicitly modeling 3D temporal dynamics. We further ablate how the quality of Vid3D depends on the number of views generated per frame. While we observe some degradation with fewer views, performance degradation remains minor. Our results thus suggest that 3D temporal knowledge may not be necessary to generate high-quality dynamic 3D scenes, potentially enabling simpler generative algorithms for this task.
Authors: Zhenzhong Kuang, Xiaochen Yang, Yingjie Shen, Chao Hu, Jun Yu
Abstract: The unprecedented capture and application of face images raise increasing concerns on anonymization to fight against privacy disclosure. Most existing methods may suffer from the problem of excessive change of the identity-independent information or insufficient identity protection. In this paper, we present a new face anonymization approach by distracting the intrinsic and extrinsic identity attentions. On the one hand, we anonymize the identity information in the feature space by distracting the intrinsic identity attention. On the other, we anonymize the visual clues (i.e. appearance and geometry structure) by distracting the extrinsic identity attention. Our approach allows for flexible and intuitive manipulation of face appearance and geometry structure to produce diverse results, and it can also be used to instruct users to perform personalized anonymization. We conduct extensive experiments on multiple datasets and demonstrate that our approach outperforms state-of-the-art methods.
Authors: Ahmad Mohammadshirazi, Ali Nosrati Firoozsalari, Mengxi Zhou, Dheeraj Kulshrestha, Rajiv Ramnath
Abstract: Automating the annotation of scanned documents is challenging, requiring a balance between computational efficiency and accuracy. DocParseNet addresses this by combining deep learning and multi-modal learning to process both text and visual data. This model goes beyond traditional OCR and semantic segmentation, capturing the interplay between text and images to preserve contextual nuances in complex document structures. Our evaluations show that DocParseNet significantly outperforms conventional models, achieving mIoU scores of 49.12 on validation and 49.78 on the test set. This reflects a 58% accuracy improvement over state-of-the-art baseline models and an 18% gain compared to the UNext baseline. Remarkably, DocParseNet achieves these results with only 2.8 million parameters, reducing the model size by approximately 25 times and speeding up training by 5 times compared to other models. These metrics, coupled with a computational efficiency of 0.034 TFLOPs (BS=1), highlight DocParseNet's high performance in document annotation. The model's adaptability and scalability make it well-suited for real-world corporate document processing applications. The code is available at https://github.com/ahmad-shirazi/DocParseNet
Authors: Younghyun Kim, Geunmin Hwang, Eunbyung Park
Abstract: Recent surge in large-scale generative models has spurred the development of vast fields in computer vision. In particular, text-to-image diffusion models have garnered widespread adoption across diverse domain due to their potential for high-fidelity image generation. Nonetheless, existing large-scale diffusion models are confined to generate images of up to 1K resolution, which is far from meeting the demands of contemporary commercial applications. Directly sampling higher-resolution images often yields results marred by artifacts such as object repetition and distorted shapes. Addressing the aforementioned issues typically necessitates training or fine-tuning models on higher resolution datasets. However, this undertaking poses a formidable challenge due to the difficulty in collecting large-scale high-resolution contents and substantial computational resources. While several preceding works have proposed alternatives, they often fail to produce convincing results. In this work, we probe the generative ability of diffusion models at higher resolution beyond its original capability and propose a novel progressive approach that fully utilizes generated low-resolution image to guide the generation of higher resolution image. Our method obviates the need for additional training or fine-tuning which significantly lowers the burden of computational costs. Extensive experiments and results validate the efficiency and efficacy of our method. Project page: https://yhyun225.github.io/DiffusHigh/
Authors: Yingjie Tang, Zixuan Li, Milo\v{s} Ha\v{s}an, Jian Yang, Beibei Wang
Abstract: Digitizing woven fabrics would be valuable for many applications, from digital humans to interior design. Previous work introduces a lightweight woven fabric acquisition approach by capturing a single reflection image and estimating the fabric parameters with a differentiable geometric and shading model. The renderings of the estimated fabric parameters can closely match the photo; however, the captured reflection image is insufficient to fully characterize the fabric sample reflectance. For instance, fabrics with different thicknesses might have similar reflection images but lead to significantly different transmission. We propose to recover the woven fabric parameters from two captured images: reflection and transmission. At the core of our method is a differentiable bidirectional scattering distribution function (BSDF) model, handling reflection and transmission, including single and multiple scattering. We propose a two-layer model, where the single scattering uses an SGGX phase function as in previous work, and multiple scattering uses a new azimuthally-invariant microflake definition, which we term ASGGX. This new fabric BSDF model closely matches real woven fabrics in both reflection and transmission. We use a simple setup for capturing reflection and transmission photos with a cell phone camera and two point lights, and estimate the fabric parameters via a lightweight network, together with a differentiable optimization. We also model the out-of-focus effects explicitly with a simple solution to match the thin-lens camera better. As a result, the renderings of the estimated parameters can agree with the input images on both reflection and transmission for the first time. The code for this paper is at https://github.com/lxtyin/FabricBTDF-Recovery.
Authors: Peiliang Zhang, Yujia Tong, Chenghu Du, Chao Che, Yongjun Zhu
Abstract: Deep learning-based RGB caries detection improves the efficiency of caries identification and is crucial for preventing oral diseases. The performance of deep learning models depends on high-quality data and requires substantial training resources, making efficient deployment challenging. Core data selection, by eliminating low-quality and confusing data, aims to enhance training efficiency without significantly compromising model performance. However, distance-based data selection methods struggle to distinguish dependencies among high-dimensional caries data. To address this issue, we propose a Core Data Selection Method with Jensen-Shannon Divergence (JSCDS) for efficient caries image learning and caries classification. We describe the core data selection criterion as the distribution of samples in different classes. JSCDS calculates the cluster centers by sample embedding representation in the caries classification network and utilizes Jensen-Shannon Divergence to compute the mutual information between data samples and cluster centers, capturing nonlinear dependencies among high-dimensional data. The average mutual information is calculated to fit the above distribution, serving as the criterion for constructing the core set for model training. Extensive experiments on RGB caries datasets show that JSCDS outperforms other data selection methods in prediction performance and time consumption. Notably, JSCDS exceeds the performance of the full dataset model with only 50% of the core data, with its performance advantage becoming more pronounced in the 70% of core data.
Authors: Huawei Sun, Hao Feng, Julius Ott, Lorenzo Servadei, Robert Wille
Abstract: Depth estimation is critical in autonomous driving for interpreting 3D scenes accurately. Recently, radar-camera depth estimation has become of sufficient interest due to the robustness and low-cost properties of radar. Thus, this paper introduces a two-stage, end-to-end trainable Confidence-aware Fusion Net (CaFNet) for dense depth estimation, combining RGB imagery with sparse and noisy radar point cloud data. The first stage addresses radar-specific challenges, such as ambiguous elevation and noisy measurements, by predicting a radar confidence map and a preliminary coarse depth map. A novel approach is presented for generating the ground truth for the confidence map, which involves associating each radar point with its corresponding object to identify potential projection surfaces. These maps, together with the initial radar input, are processed by a second encoder. For the final depth estimation, we innovate a confidence-aware gated fusion mechanism to integrate radar and image features effectively, thereby enhancing the reliability of the depth map by filtering out radar noise. Our methodology, evaluated on the nuScenes dataset, demonstrates superior performance, improving upon the current leading model by 3.2% in Mean Absolute Error (MAE) and 2.7% in Root Mean Square Error (RMSE). Code: https://github.com/harborsarah/CaFNet
Authors: Kalibinuer Tiliwalidi, Chengyin Hu, Weiwen Shi
Abstract: While extensive research exists on physical adversarial attacks within the visible spectrum, studies on such techniques in the infrared spectrum are limited. Infrared object detectors are vital in modern technological applications but are susceptible to adversarial attacks, posing significant security threats. Previous studies using physical perturbations like light bulb arrays and aerogels for white-box attacks, or hot and cold patches for black-box attacks, have proven impractical or limited in multi-view support. To address these issues, we propose the Adversarial Infrared Grid (AdvGrid), which models perturbations in a grid format and uses a genetic algorithm for black-box optimization. These perturbations are cyclically applied to various parts of a pedestrian's clothing to facilitate multi-view black-box physical attacks on infrared pedestrian detectors. Extensive experiments validate AdvGrid's effectiveness, stealthiness, and robustness. The method achieves attack success rates of 80.00\% in digital environments and 91.86\% in physical environments, outperforming baseline methods. Additionally, the average attack success rate exceeds 50\% against mainstream detectors, demonstrating AdvGrid's robustness. Our analyses include ablation studies, transfer attacks, and adversarial defenses, confirming the method's superiority.
Authors: Avraham Raviv, Yizhak Y. Elboher, Michelle Aluf-Medina, Yael Leibovich Weiss, Omer Cohen, Roy Assa, Guy Katz, Hillel Kugler
Abstract: Deep Neural Networks (DNNs) are ubiquitous in real-world applications, yet they remain vulnerable to errors and adversarial attacks. This work tackles the challenge of applying formal verification to ensure the safety of computer vision models, extending verification beyond image classification to object detection. We propose a general formulation for certifying the robustness of object detection models using formal verification and outline implementation strategies compatible with state-of-the-art verification tools. Our approach enables the application of these tools, originally designed for verifying classification models, to object detection. We define various attacks for object detection, illustrating the diverse ways adversarial inputs can compromise neural network outputs. Our experiments, conducted on several common datasets and networks, reveal potential errors in object detection models, highlighting system vulnerabilities and emphasizing the need for expanding formal verification to these new domains. This work paves the way for further research in integrating formal verification across a broader range of computer vision applications.
Authors: Zixing Li, Chao Yan, Zhen Lan, Xiaojia Xiang, Han Zhou, Jun Lai, Dengqing Tang
Abstract: Advanced cognition can be extracted from the human brain using brain-computer interfaces. Integrating these interfaces with computer vision techniques, which possess efficient feature extraction capabilities, can achieve more robust and accurate detection of dim targets in aerial images. However, existing target detection methods primarily concentrate on homogeneous data, lacking efficient and versatile processing capabilities for heterogeneous multimodal data. In this paper, we first build a brain-eye-computer based object detection system for aerial images under few-shot conditions. This system detects suspicious targets using region proposal networks, evokes the event-related potential (ERP) signal in electroencephalogram (EEG) through the eye-tracking-based slow serial visual presentation (ESSVP) paradigm, and constructs the EEG-image data pairs with eye movement data. Then, an adaptive modality balanced online knowledge distillation (AMBOKD) method is proposed to recognize dim objects with the EEG-image data. AMBOKD fuses EEG and image features using a multi-head attention module, establishing a new modality with comprehensive features. To enhance the performance and robust capability of the fusion modality, simultaneous training and mutual learning between modalities are enabled by end-to-end online knowledge distillation. During the learning process, an adaptive modality balancing module is proposed to ensure multimodal equilibrium by dynamically adjusting the weights of the importance and the training gradients across various modalities. The effectiveness and superiority of our method are demonstrated by comparing it with existing state-of-the-art methods. Additionally, experiments conducted on public datasets and system validations in real-world scenarios demonstrate the reliability and practicality of the proposed system and the designed method.
Authors: Kaixin Xu, Zhe Wang, Chunyun Chen, Xue Geng, Jie Lin, Xulei Yang, Min Wu, Xiaoli Li, Weisi Lin
Abstract: Vision transformers have emerged as a promising alternative to convolutional neural networks for various image analysis tasks, offering comparable or superior performance. However, one significant drawback of ViTs is their resource-intensive nature, leading to increased memory footprint, computation complexity, and power consumption. To democratize this high-performance technology and make it more environmentally friendly, it is essential to compress ViT models, reducing their resource requirements while maintaining high performance. In this paper, we introduce a new block-structured pruning to address the resource-intensive issue for ViTs, offering a balanced trade-off between accuracy and hardware acceleration. Unlike unstructured pruning or channel-wise structured pruning, block pruning leverages the block-wise structure of linear layers, resulting in more efficient matrix multiplications. To optimize this pruning scheme, our paper proposes a novel hardware-aware learning objective that simultaneously maximizes speedup and minimizes power consumption during inference, tailored to the block sparsity structure. This objective eliminates the need for empirical look-up tables and focuses solely on reducing parametrized layer connections. Moreover, our paper provides a lightweight algorithm to achieve post-training pruning for ViTs, utilizing second-order Taylor approximation and empirical optimization to solve the proposed hardware-aware objective. Extensive experiments on ImageNet are conducted across various ViT architectures, including DeiT-B and DeiT-S, demonstrating competitive performance with other pruning methods and achieving a remarkable balance between accuracy preservation and power savings. Especially, we achieve up to 3.93x and 1.79x speedups on dedicated hardware and GPUs respectively for DeiT-B, and also observe an inference power reduction by 1.4x on real-world GPUs.
Authors: Bohan Li, Jiajun Deng, Wenyao Zhang, Zhujin Liang, Dalong Du, Xin Jin, Wenjun Zeng
Abstract: Camera-based 3D semantic scene completion (SSC) is pivotal for predicting complicated 3D layouts with limited 2D image observations. The existing mainstream solutions generally leverage temporal information by roughly stacking history frames to supplement the current frame, such straightforward temporal modeling inevitably diminishes valid clues and increases learning difficulty. To address this problem, we present HTCL, a novel Hierarchical Temporal Context Learning paradigm for improving camera-based semantic scene completion. The primary innovation of this work involves decomposing temporal context learning into two hierarchical steps: (a) cross-frame affinity measurement and (b) affinity-based dynamic refinement. Firstly, to separate critical relevant context from redundant information, we introduce the pattern affinity with scale-aware isolation and multiple independent learners for fine-grained contextual correspondence modeling. Subsequently, to dynamically compensate for incomplete observations, we adaptively refine the feature sampling locations based on initially identified locations with high affinity and their neighboring relevant regions. Our method ranks $1^{st}$ on the SemanticKITTI benchmark and even surpasses LiDAR-based methods in terms of mIoU on the OpenOccupancy benchmark. Our code is available on https://github.com/Arlo0o/HTCL.
Authors: Tinghuai Wang, Guangming Wang, Kuan Eeik Tan
Abstract: Convolutional neural networks (CNN) have made significant advances in detecting roads from satellite images. However, existing CNN approaches are generally repurposed semantic segmentation architectures and suffer from the poor delineation of long and curved regions. Lack of overall road topology and structure information further deteriorates their performance on challenging remote sensing images. This paper presents a novel multi-task graph neural network (GNN) which simultaneously detects both road regions and road borders; the inter-play between these two tasks unlocks superior performance from two perspectives: (1) the hierarchically detected road borders enable the network to capture and encode holistic road structure to enhance road connectivity (2) identifying the intrinsic correlation of semantic landcover regions mitigates the difficulty in recognizing roads cluttered by regions with similar appearance. Experiments on challenging dataset demonstrate that the proposed architecture can improve the road border delineation and road extraction accuracy compared with the existing methods.
Authors: Huiling Wang
Abstract: Convolutional neural networks (CNNs) have achieved remarkable performance in hyperspectral image (HSI) classification over the last few years. Despite the progress that has been made, rich and informative spectral information of HSI has been largely underutilized by existing methods which employ convolutional kernels with limited size of receptive field in the spectral domain. To address this issue, we propose a spectral graph reasoning network (SGR) learning framework comprising two crucial modules: 1) a spectral decoupling module which unpacks and casts multiple spectral embeddings into a unified graph whose node corresponds to an individual spectral feature channel in the embedding space; the graph performs interpretable reasoning to aggregate and align spectral information to guide learning spectral-specific graph embeddings at multiple contextual levels 2) a spectral ensembling module explores the interactions and interdependencies across graph embedding hierarchy via a novel recurrent graph propagation mechanism. Experiments on two HSI datasets demonstrate that the proposed architecture can significantly improve the classification accuracy compared with the existing methods with a sizable margin.
Authors: Ruizhuo Song, Beiming Yuan
Abstract: Visual abstract reasoning problems hold immense importance in the field of image processing. Both Bongard-Logo and Raven's Progressive Matrices (RPM) belong to this domain, with Bongard-Logo categorized as image clustering reasoning and RPM involving image progression pattern reasoning. This paper introduces Valen, a novel baseline model under probabilistic highlighting models. Valen exhibits remarkable performance in solving both RPM and Bongard-Logo problems, offering a versatile solution. Our investigation delves into the underlying mechanisms of probability-highlighting solvers, realizing they approximate solutions to reasoning problem instances as distributions delineated by primary and auxiliary samples. We propose that the learning objective is not the distribution of correct solutions but one defined by both primary and auxiliary samples. To bridge discrepancies, we introduced the Tine method, an adversarial learning-based approach to assist Valen in estimating a solution distribution closer to the correct one, albeit with issues like unstable training. Reflecting on Tine, we propose modeling the sample distribution of reasoning problems as a mixture of Gaussian distributions, leading to the Funny method. This effectively enables Valen to capture the true form of the correct solution distribution. Furthermore, we designed the SBR method to model the distribution of progressive patterns representation similarly. Overall, the Funny, Tine, and SBR methods significantly improve Valen's performance, providing new ideas and methods for studying visual abstract reasoning problems.
Authors: Weitai Kang, Gaowen Liu, Mubarak Shah, Yan Yan
Abstract: Different from Object Detection, Visual Grounding deals with detecting a bounding box for each text-image pair. This one box for each text-image data provides sparse supervision signals. Although previous works achieve impressive results, their passive utilization of annotation, i.e. the sole use of the box annotation as regression ground truth, results in a suboptimal performance. In this paper, we present SegVG, a novel method transfers the box-level annotation as Segmentation signals to provide an additional pixel-level supervision for Visual Grounding. Specifically, we propose the Multi-layer Multi-task Encoder-Decoder as the target grounding stage, where we learn a regression query and multiple segmentation queries to ground the target by regression and segmentation of the box in each decoding layer, respectively. This approach allows us to iteratively exploit the annotation as signals for both box-level regression and pixel-level segmentation. Moreover, as the backbones are typically initialized by pretrained parameters learned from unimodal tasks and the queries for both regression and segmentation are static learnable embeddings, a domain discrepancy remains among these three types of features, which impairs subsequent target grounding. To mitigate this discrepancy, we introduce the Triple Alignment module, where the query, text, and vision tokens are triangularly updated to share the same space by triple attention mechanism. Extensive experiments on five widely used datasets validate our state-of-the-art (SOTA) performance.
Authors: Weitai Kang, Luowei Zhou, Junyi Wu, Changchang Sun, Yan Yan
Abstract: Unlike Object Detection, Visual Grounding task necessitates the detection of an object described by complex free-form language. To simultaneously model such complex semantic and visual representations, recent state-of-the-art studies adopt transformer-based models to fuse features from both modalities, further introducing various modules that modulate visual features to align with the language expressions and eliminate the irrelevant redundant information. However, their loss function, still adopting common Object Detection losses, solely governs the bounding box regression output, failing to fully optimize for the above objectives. To tackle this problem, in this paper, we first analyze the attention mechanisms of transformer-based models. Building upon this, we further propose a novel framework named Attention-Driven Constraint Balancing (AttBalance) to optimize the behavior of visual features within language-relevant regions. Extensive experimental results show that our method brings impressive improvements. Specifically, we achieve constant improvements over five different models evaluated on four different benchmarks. Moreover, we attain a new state-of-the-art performance by integrating our method into QRNet.
Authors: Weitai Kang, Mengxue Qu, Yunchao Wei, Yan Yan
Abstract: Semi-Supervised Visual Grounding (SSVG) is a new challenge for its sparse labeled data with the need for multimodel understanding. A previous study, RefTeacher, makes the first attempt to tackle this task by adopting the teacher-student framework to provide pseudo confidence supervision and attention-based supervision. However, this approach is incompatible with current state-of-the-art visual grounding models, which follow the Transformer-based pipeline. These pipelines directly regress results without region proposals or foreground binary classification, rendering them unsuitable for fitting in RefTeacher due to the absence of confidence scores. Furthermore, the geometric difference in teacher and student inputs, stemming from different data augmentations, induces natural misalignment in attention-based constraints. To establish a compatible SSVG framework, our paper proposes the ACTive REtraining approach for Semi-Supervised Visual Grounding, abbreviated as ACTRESS. Initially, the model is enhanced by incorporating an additional quantized detection head to expose its detection confidence. Building upon this, ACTRESS consists of an active sampling strategy and a selective retraining strategy. The active sampling strategy iteratively selects high-quality pseudo labels by evaluating three crucial aspects: Faithfulness, Robustness, and Confidence, optimizing the utilization of unlabeled data. The selective retraining strategy retrains the model with periodic re-initialization of specific parameters, facilitating the model's escape from local minima. Extensive experiments demonstrates our superior performance on widely-used benchmark datasets.
Authors: Aleksandar Vorkapic, Miran Pobar, Marina Ivasic-Kos
Abstract: This research presents a novel application of computer vision (CV) and deep learning methods for real-time sea state recognition, aiming to contribute to improving the operational safety and energy efficiency of seagoing vessels, key factors in meeting the legislative carbon reduction targets. Our work focuses on utilizing sea images in operational envelopes captured by a single stationary camera mounted on the ship bridge. The collected images are used to train a deep learning model to automatically recognize the state of the sea based on the Beaufort scale. To recognize the sea state, we used 4 state-of-the-art deep neural networks with different characteristics that proved useful in various computer vision tasks: Resnet-101, NASNet, MobileNet_v2, and Transformer ViT-b32. Furthermore, we have defined a unique large-scale dataset, collected over a broad range of sea conditions from an ocean-going vessel prepared for machine learning. We used the transfer learning approach to fine-tune the models on our dataset. The obtained results demonstrate the potential for this approach to complement traditional methods, particularly where in-situ measurements are unfeasible or interpolated weather buoy data is insufficiently accurate. This study sets the groundwork for further development of sea state classification models to address recognized gaps in maritime research and enable safer and more efficient maritime operations.
Authors: Dimitrios Kollias, Stefanos Zafeiriou, Irene Kotsia, Abhinav Dhall, Shreya Ghosh, Chunchang Shao, Guanyu Hu
Abstract: This paper describes the 7th Affective Behavior Analysis in-the-wild (ABAW) Competition, which is part of the respective Workshop held in conjunction with ECCV 2024. The 7th ABAW Competition addresses novel challenges in understanding human expressions and behaviors, crucial for the development of human-centered technologies. The Competition comprises of two sub-challenges: i) Multi-Task Learning (the goal is to learn at the same time, in a multi-task learning setting, to estimate two continuous affect dimensions, valence and arousal, to recognise between the mutually exclusive classes of the 7 basic expressions and 'other'), and to detect 12 Action Units); and ii) Compound Expression Recognition (the target is to recognise between the 7 mutually exclusive compound expression classes). s-Aff-Wild2, which is a static version of the A/V Aff-Wild2 database and contains annotations for valence-arousal, expressions and Action Units, is utilized for the purposes of the Multi-Task Learning Challenge; a part of C-EXPR-DB, which is an A/V in-the-wild database with compound expression annotations, is utilized for the purposes of the Compound Expression Recognition Challenge. In this paper, we introduce the two challenges, detailing their datasets and the protocols followed for each. We also outline the evaluation metrics, and highlight the baseline systems and their results. Additional information about the competition can be found at \url{https://affective-behavior-analysis-in-the-wild.github.io/7th}.
URLs: https://affective-behavior-analysis-in-the-wild.github.io/7th
Authors: Ruixiao Zhang, Yihong Wu, Juheon Lee, Adam Prugel-Bennett, Xiaohao Cai
Abstract: The performance of domain adaptation technologies has not yet reached an ideal level in the current 3D object detection field for autonomous driving, which is mainly due to significant differences in the size of vehicles, as well as the environments they operate in when applied across domains. These factors together hinder the effective transfer and application of knowledge learned from specific datasets. Since the existing evaluation metrics are initially designed for evaluation on a single domain by calculating the 2D or 3D overlap between the prediction and ground-truth bounding boxes, they often suffer from the overfitting problem caused by the size differences among datasets. This raises a fundamental question related to the evaluation of the 3D object detection models' cross-domain performance: Do we really need models to maintain excellent performance in their original 3D bounding boxes after being applied across domains? From a practical application perspective, one of our main focuses is actually on preventing collisions between vehicles and other obstacles, especially in cross-domain scenarios where correctly predicting the size of vehicles is much more difficult. In other words, as long as a model can accurately identify the closest surfaces to the ego vehicle, it is sufficient to effectively avoid obstacles. In this paper, we propose two metrics to measure 3D object detection models' ability of detecting the closer surfaces to the sensor on the ego vehicle, which can be used to evaluate their cross-domain performance more comprehensively and reasonably. Furthermore, we propose a refinement head, named EdgeHead, to guide models to focus more on the learnable closer surfaces, which can greatly improve the cross-domain performance of existing models not only under our new metrics, but even also under the original BEV/3D metrics.
Authors: Alex Costanzino, Pierluigi Zama Ramirez, Giuseppe Lisanti, Luigi Di Stefano
Abstract: Motivated by efficiency requirements, most anomaly detection and segmentation (AD&S) methods focus on processing low-resolution images, e.g., $224\times 224$ pixels, obtained by downsampling the original input images. In this setting, downsampling is typically applied also to the provided ground-truth defect masks. Yet, as numerous industrial applications demand identification of both large and tiny defects, the above-described protocol may fall short in providing a realistic picture of the actual performance attainable by current methods. Hence, in this work, we introduce a novel benchmark that evaluates methods on the original, high-resolution image and ground-truth masks, focusing on segmentation performance as a function of the size of anomalies. Our benchmark includes a metric that captures robustness with respect to defect size, i.e., the ability of a method to preserve good localization from large anomalies to tiny ones. Furthermore, we introduce an AD&S approach based on a novel Teacher-Student paradigm which relies on two shallow MLPs (the Students) that learn to transfer patch features across the layers of a frozen vision transformer (the Teacher). By means of our benchmark, we evaluate our proposal and other recent AD&S methods on high-resolution inputs containing large and tiny defects. Our proposal features the highest robustness to defect size, runs at the fastest speed, yields state-of-the-art performance on the MVTec AD dataset and state-of-the-art segmentation performance on the VisA dataset.
Authors: Griffin Golias, Masa Nakura-Fan, Vitaly Ablavsky
Abstract: We propose a graph-based tracking formulation for multi-object tracking (MOT) where target detections contain kinematic information and re-identification features (attributes). Our method applies a successive shortest paths (SSP) algorithm to a tracking graph defined over a batch of frames. The edge costs in this tracking graph are computed via a message-passing network, a graph neural network (GNN) variant. The parameters of the GNN, and hence, the tracker, are learned end-to-end on a training set of example ground-truth tracks and detections. Specifically, learning takes the form of bilevel optimization guided by our novel loss function. We evaluate our algorithm on simulated scenarios to understand its sensitivity to scenario aspects and model hyperparameters. Across varied scenario complexities, our method compares favorably to a strong baseline.
Authors: Ananthu Aniraj, Cassio F. Dantas, Dino Ienco, Diego Marcos
Abstract: Computer vision methods that explicitly detect object parts and reason on them are a step towards inherently interpretable models. Existing approaches that perform part discovery driven by a fine-grained classification task make very restrictive assumptions on the geometric properties of the discovered parts; they should be small and compact. Although this prior is useful in some cases, in this paper we show that pre-trained transformer-based vision models, such as self-supervised DINOv2 ViT, enable the relaxation of these constraints. In particular, we find that a total variation (TV) prior, which allows for multiple connected components of any size, substantially outperforms previous work. We test our approach on three fine-grained classification benchmarks: CUB, PartImageNet and Oxford Flowers, and compare our results to previously published methods as well as a re-implementation of the state-of-the-art method PDiscoNet with a transformer-based backbone. We consistently obtain substantial improvements across the board, both on part discovery metrics and the downstream classification task, showing that the strong inductive biases in self-supervised ViT models require to rethink the geometric priors that can be used for unsupervised part discovery.
Authors: Kam Woh Ng, Xiatian Zhu, Yi-Zhe Song, Tao Xiang
Abstract: This paper propels creative control in generative visual AI by allowing users to "select". Departing from traditional text or sketch-based methods, we for the first time allow users to choose visual concepts by parts for their creative endeavors. The outcome is fine-grained generation that precisely captures selected visual concepts, ensuring a holistically faithful and plausible result. To achieve this, we first parse objects into parts through unsupervised feature clustering. Then, we encode parts into text tokens and introduce an entropy-based normalized attention loss that operates on them. This loss design enables our model to learn generic prior topology knowledge about object's part composition, and further generalize to novel part compositions to ensure the generation looks holistically faithful. Lastly, we employ a bottleneck encoder to project the part tokens. This not only enhances fidelity but also accelerates learning, by leveraging shared knowledge and facilitating information exchange among instances. Visual results in the paper and supplementary material showcase the compelling power of PartCraft in crafting highly customized, innovative creations, exemplified by the "charming" and creative birds. Code is released at https://github.com/kamwoh/partcraft.
Authors: Yu Guo, Yuan Gao, Yuxu Lu, Huilin Zhu, Ryan Wen Liu, Shengfeng He
Abstract: In real-world scenarios, image impairments often manifest as composite degradations, presenting a complex interplay of elements such as low light, haze, rain, and snow. Despite this reality, existing restoration methods typically target isolated degradation types, thereby falling short in environments where multiple degrading factors coexist. To bridge this gap, our study proposes a versatile imaging model that consolidates four physical corruption paradigms to accurately represent complex, composite degradation scenarios. In this context, we propose OneRestore, a novel transformer-based framework designed for adaptive, controllable scene restoration. The proposed framework leverages a unique cross-attention mechanism, merging degraded scene descriptors with image features, allowing for nuanced restoration. Our model allows versatile input scene descriptors, ranging from manual text embeddings to automatic extractions based on visual attributes. Our methodology is further enhanced through a composite degradation restoration loss, using extra degraded images as negative samples to fortify model constraints. Comparative results on synthetic and real-world datasets demonstrate OneRestore as a superior solution, significantly advancing the state-of-the-art in addressing complex, composite degradations.
Authors: Xu Song, Tianyu Shen, Hui Li, Xiao-Jun Wu
Abstract: In image fusion tasks, an ideal image decomposition method can bring better performance. MDLatLRR has done a great job in this aspect, but there is still exist some space for improvement. Considering that MDLatLRR focuses solely on the detailed parts (salient features) extracted from input images via latent low-rank representation (LatLRR), the basic parts (principal features) extracted by LatLRR are not fully utilized. Therefore, we introduced an enhanced multi-level decomposition method named dual-decomposed MDLatLRR (D2-LRR) which effectively analyzes and utilizes all image features extracted through LatLRR. Specifically, color images are converted into YUV color space and grayscale images, and the Y-channel and grayscale images are input into the trained parameters of LatLRR to obtain the detailed parts containing four rounds of decomposition and the basic parts. Subsequently, the basic parts are fused using an average strategy, while the detail part is fused using kernel norm operation. The fused image is ultimately transformed back into an RGB image, resulting in the final fusion output. We apply D2-LRR to medical image fusion tasks. The detailed parts are fused employing a nuclear-norm operation, while the basic parts are fused using an average strategy. Comparative analyses among existing methods showcase that our proposed approach attains cutting-edge fusion performance in both objective and subjective assessments.
Authors: Lukas Christ, Shahin Amiriparian, Alexander Kathan, Niklas M\"uller, Andreas K\"onig, Bj\"orn W. Schuller
Abstract: Humor is a substantial element of human social behavior, affect, and cognition. Its automatic understanding can facilitate a more naturalistic human-AI interaction. Current methods of humor detection have been exclusively based on staged data, making them inadequate for "real-world" applications. We contribute to addressing this deficiency by introducing the novel Passau-Spontaneous Football Coach Humor (Passau-SFCH) dataset, comprising about 11 hours of recordings. The Passau-SFCH dataset is annotated for the presence of humor and its dimensions (sentiment and direction) as proposed in Martin's Humor Style Questionnaire. We conduct a series of experiments employing pretrained Transformers, convolutional neural networks, and expert-designed features. The performance of each modality (text, audio, video) for spontaneous humor recognition is analyzed and their complementarity is investigated. Our findings suggest that for the automatic analysis of humor and its sentiment, facial expressions are most promising, while humor direction can be best modeled via text-based features. Further, we experiment with different multimodal approaches to humor recognition, including decision-level fusion and MulT, a multimodal Transformer approach. In this context, we propose a novel multimodal architecture that yields the best overall results. Finally, we make our code publicly available at https://www.github.com/lc0197/passau-sfch. The Passau-SFCH dataset is available upon request.
Authors: Muhammad Usman Akbar, Wuhao Wang, Anders Eklund
Abstract: Diffusion models were initially developed for text-to-image generation and are now being utilized to generate high quality synthetic images. Preceded by GANs, diffusion models have shown impressive results using various evaluation metrics. However, commonly used metrics such as FID and IS are not suitable for determining whether diffusion models are simply reproducing the training images. Here we train StyleGAN and a diffusion model, using BRATS20, BRATS21 and a chest x-ray pneumonia dataset, to synthesize brain MRI and chest x-ray images, and measure the correlation between the synthetic images and all training images. Our results show that diffusion models are more likely to memorize the training images, compared to StyleGAN, especially for small datasets and when using 2D slices from 3D volumes. Researchers should be careful when using diffusion models (and to some extent GANs) for medical imaging, if the final goal is to share the synthetic images.
Authors: Subba Reddy Oota, Zijiao Chen, Manish Gupta, Raju S. Bapi, Gael Jobard, Frederic Alexandre, Xavier Hinaut
Abstract: Can we obtain insights about the brain using AI models? How is the information in deep learning models related to brain recordings? Can we improve AI models with the help of brain recordings? Such questions can be tackled by studying brain recordings like functional magnetic resonance imaging (fMRI). As a first step, the neuroscience community has contributed several large cognitive neuroscience datasets related to passive reading/listening/viewing of concept words, narratives, pictures, and movies. Encoding and decoding models using these datasets have also been proposed in the past two decades. These models serve as additional tools for basic cognitive science and neuroscience research. Encoding models aim at generating fMRI brain representations given a stimulus automatically. They have several practical applications in evaluating and diagnosing neurological conditions and thus may also help design therapies for brain damage. Decoding models solve the inverse problem of reconstructing the stimuli given the fMRI. They are useful for designing brain-machine or brain-computer interfaces. Inspired by the effectiveness of deep learning models for natural language processing, computer vision, and speech, several neural encoding and decoding models have been recently proposed. In this survey, we will first discuss popular representations of language, vision and speech stimuli, and present a summary of neuroscience datasets. Further, we will review popular deep learning based encoding and decoding architectures and note their benefits and limitations. Finally, we will conclude with a summary and discussion about future trends. Given the large amount of recently published work in the computational cognitive neuroscience (CCN) community, we believe that this survey enables an entry point for DNN researchers to diversify into CCN research.
Authors: Karim Lekadir, Aasa Feragen, Abdul Joseph Fofanah, Alejandro F Frangi, Alena Buyx, Anais Emelie, Andrea Lara, Antonio R Porras, An-Wen Chan, Arcadi Navarro, Ben Glocker, Benard O Botwe, Bishesh Khanal, Brigit Beger, Carol C Wu, Celia Cintas, Curtis P Langlotz, Daniel Rueckert, Deogratias Mzurikwao, Dimitrios I Fotiadis, Doszhan Zhussupov, Enzo Ferrante, Erik Meijering, Eva Weicken, Fabio A Gonz\'alez, Folkert W Asselbergs, Fred Prior, Gabriel P Krestin, Gary Collins, Geletaw S Tegenaw, Georgios Kaissis, Gianluca Misuraca, Gianna Tsakou, Girish Dwivedi, Haridimos Kondylakis, Harsha Jayakody, Henry C Woodruf, Horst Joachim Mayer, Hugo JWL Aerts, Ian Walsh, Ioanna Chouvarda, Ir\`ene Buvat, Isabell Tributsch, Islem Rekik, James Duncan, Jayashree Kalpathy-Cramer, Jihad Zahir, Jinah Park, John Mongan, Judy W Gichoya, Julia A Schnabel, Kaisar Kushibar, Katrine Riklund, Kensaku Mori, Kostas Marias, Lameck M Amugongo, Lauren A Fromont, Lena Maier-Hein, Leonor Cerd\'a Alberich, Leticia Rittner, Lighton Phiri, Linda Marrakchi-Kacem, Llu\'is Donoso-Bach, Luis Mart\'i-Bonmat\'i, M Jorge Cardoso, Maciej Bobowicz, Mahsa Shabani, Manolis Tsiknakis, Maria A Zuluaga, Maria Bielikova, Marie-Christine Fritzsche, Marina Camacho, Marius George Linguraru, Markus Wenzel, Marleen De Bruijne, Martin G Tolsgaard, Marzyeh Ghassemi, Md Ashrafuzzaman, Melanie Goisauf, Mohammad Yaqub, M\'onica Cano Abad\'ia, Mukhtar M E Mahmoud, Mustafa Elattar, Nicola Rieke, Nikolaos Papanikolaou, Noussair Lazrak, Oliver D\'iaz, Olivier Salvado, Oriol Pujol, Ousmane Sall, Pamela Guevara, Peter Gordebeke, Philippe Lambin, Pieta Brown, Purang Abolmaesumi, Qi Dou, Qinghua Lu, Richard Osuala, Rose Nakasi, S Kevin Zhou, Sandy Napel, Sara Colantonio, Shadi Albarqouni, Smriti Joshi, Stacy Carter, Stefan Klein, Steffen E Petersen, Susanna Auss\'o, Suyash Awate, Tammy Riklin Raviv, Tessa Cook, Tinashe E M Mutsvangwa, Wendy A Rogers, Wiro J Niessen, X\`enia Puig-Bosch, Yi Zeng, Yunusa G Mohammed, Yves Saint James Aquino, Zohaib Salahuddin, Martijn P A Starmans
Abstract: Despite major advances in artificial intelligence (AI) for medicine and healthcare, the deployment and adoption of AI technologies remain limited in real-world clinical practice. In recent years, concerns have been raised about the technical, clinical, ethical and legal risks associated with medical AI. To increase real world adoption, it is essential that medical AI tools are trusted and accepted by patients, clinicians, health organisations and authorities. This work describes the FUTURE-AI guideline as the first international consensus framework for guiding the development and deployment of trustworthy AI tools in healthcare. The FUTURE-AI consortium was founded in 2021 and currently comprises 118 inter-disciplinary experts from 51 countries representing all continents, including AI scientists, clinicians, ethicists, and social scientists. Over a two-year period, the consortium defined guiding principles and best practices for trustworthy AI through an iterative process comprising an in-depth literature review, a modified Delphi survey, and online consensus meetings. The FUTURE-AI framework was established based on 6 guiding principles for trustworthy AI in healthcare, i.e. Fairness, Universality, Traceability, Usability, Robustness and Explainability. Through consensus, a set of 28 best practices were defined, addressing technical, clinical, legal and socio-ethical dimensions. The recommendations cover the entire lifecycle of medical AI, from design, development and validation to regulation, deployment, and monitoring. FUTURE-AI is a risk-informed, assumption-free guideline which provides a structured approach for constructing medical AI tools that will be trusted, deployed and adopted in real-world practice. Researchers are encouraged to take the recommendations into account in proof-of-concept stages to facilitate future translation towards clinical practice of medical AI.
Authors: Yang Miao, Iro Armeni, Marc Pollefeys, Daniel Barath
Abstract: We introduce an online 2D-to-3D semantic instance mapping algorithm aimed at generating comprehensive, accurate, and efficient semantic 3D maps suitable for autonomous agents in unstructured environments. The proposed approach is based on a Voxel-TSDF representation used in recent algorithms. It introduces novel ways of integrating semantic prediction confidence during mapping, producing semantic and instance-consistent 3D regions. Further improvements are achieved by graph optimization-based semantic labeling and instance refinement. The proposed method achieves accuracy superior to the state of the art on public large-scale datasets, improving on a number of widely used metrics. We also highlight a downfall in the evaluation of recent studies: using the ground truth trajectory as input instead of a SLAM-estimated one substantially affects the accuracy, creating a large gap between the reported results and the actual performance on real-world data.
Authors: Lukas Struppek, Dominik Hintersdorf, Kristian Kersting
Abstract: Label smoothing -- using softened labels instead of hard ones -- is a widely adopted regularization method for deep learning, showing diverse benefits such as enhanced generalization and calibration. Its implications for preserving model privacy, however, have remained unexplored. To fill this gap, we investigate the impact of label smoothing on model inversion attacks (MIAs), which aim to generate class-representative samples by exploiting the knowledge encoded in a classifier, thereby inferring sensitive information about its training data. Through extensive analyses, we uncover that traditional label smoothing fosters MIAs, thereby increasing a model's privacy leakage. Even more, we reveal that smoothing with negative factors counters this trend, impeding the extraction of class-related information and leading to privacy preservation, beating state-of-the-art defenses. This establishes a practical and powerful novel way for enhancing model resilience against MIAs.
Authors: Pengxiang Ding, Han Zhao, Wenxuan Song, Wenjie Zhang, Min Zhang, Siteng Huang, Ningxi Yang, Donglin Wang
Abstract: The important manifestation of robot intelligence is the ability to naturally interact and autonomously make decisions. Traditional approaches to robot control often compartmentalize perception, planning, and decision-making, simplifying system design but limiting the synergy between different information streams. This compartmentalization poses challenges in achieving seamless autonomous reasoning, decision-making, and action execution. To address these limitations, a novel paradigm, named Vision-Language-Action tasks for QUAdruped Robots (QUAR-VLA), has been introduced in this paper. This approach tightly integrates visual information and instructions to generate executable actions, effectively merging perception, planning, and decision-making. The central idea is to elevate the overall intelligence of the robot. Within this framework, a notable challenge lies in aligning fine-grained instructions with visual perception information. This emphasizes the complexity involved in ensuring that the robot accurately interprets and acts upon detailed instructions in harmony with its visual observations. Consequently, we propose QUAdruped Robotic Transformer (QUART), a family of VLA models to integrate visual information and instructions from diverse modalities as input and generates executable actions for real-world robots and present QUAdruped Robot Dataset (QUARD), a large-scale multi-task dataset including navigation, complex terrain locomotion, and whole-body manipulation tasks for training QUART models. Our extensive evaluation (4000 evaluation trials) shows that our approach leads to performant robotic policies and enables QUART to obtain a range of emergent capabilities.
Authors: Jihan Yang, Runyu Ding, Ellis Brown, Xiaojuan Qi, Saining Xie
Abstract: There is a sensory gulf between the Earth that humans inhabit and the digital realms in which modern AI agents are created. To develop AI agents that can sense, think, and act as flexibly as humans in real-world settings, it is imperative to bridge the realism gap between the digital and physical worlds. How can we embody agents in an environment as rich and diverse as the one we inhabit, without the constraints imposed by real hardware and control? Towards this end, we introduce V-IRL: a platform that enables agents to scalably interact with the real world in a virtual yet realistic environment. Our platform serves as a playground for developing agents that can accomplish various practical tasks and as a vast testbed for measuring progress in capabilities spanning perception, decision-making, and interaction with real-world data across the entire globe.
Authors: Antonia W\"ust, Wolfgang Stammer, Quentin Delfosse, Devendra Singh Dhami, Kristian Kersting
Abstract: The challenge in learning abstract concepts from images in an unsupervised fashion lies in the required integration of visual perception and generalizable relational reasoning. Moreover, the unsupervised nature of this task makes it necessary for human users to be able to understand a model's learnt concepts and potentially revise false behaviours. To tackle both the generalizability and interpretability constraints of visual concept learning, we propose Pix2Code, a framework that extends program synthesis to visual relational reasoning by utilizing the abilities of both explicit, compositional symbolic and implicit neural representations. This is achieved by retrieving object representations from images and synthesizing relational concepts as lambda-calculus programs. We evaluate the diverse properties of Pix2Code on the challenging reasoning domains, Kandinsky Patterns and CURI, thereby testing its ability to identify compositional visual concepts that generalize to novel data and concept configurations. Particularly, in stark contrast to neural approaches, we show that Pix2Code's representations remain human interpretable and can be easily revised for improved performance.
Authors: Ka Man Lo, Yiming Liang, Wenyu Du, Yuantao Fan, Zili Wang, Wenhao Huang, Lei Ma, Jie Fu
Abstract: Modular neural architectures are gaining attention for their powerful generalization and efficient adaptation to new domains. However, training these models poses challenges due to optimization difficulties arising from intrinsic sparse connectivity. Leveraging knowledge from monolithic models through techniques like knowledge distillation can facilitate training and enable integration of diverse knowledge. Nevertheless, conventional knowledge distillation approaches are not tailored to modular models and struggle with unique architectures and enormous parameter counts. Motivated by these challenges, we propose module-to-module knowledge distillation (m2mKD) for transferring knowledge between modules. m2mKD combines teacher modules of a pretrained monolithic model and student modules of a modular model with a shared meta model respectively to encourage the student module to mimic the behaviour of the teacher module. We evaluate m2mKD on two modular neural architectures: Neural Attentive Circuits (NACs) and Vision Mixture-of-Experts (V-MoE). Applying m2mKD to NACs yields significant improvements in IID accuracy on Tiny-ImageNet (up to 5.6%) and OOD robustness on Tiny-ImageNet-R (up to 4.2%). Additionally, the V-MoE-Base model trained with m2mKD achieves 3.5% higher accuracy than end-to-end training on ImageNet-1k. Code is available at https://github.com/kamanphoebe/m2mKD.
Authors: Yuanhao Cai, Yixun Liang, Jiahao Wang, Angtian Wang, Yulun Zhang, Xiaokang Yang, Zongwei Zhou, Alan Yuille
Abstract: X-ray is widely applied for transmission imaging due to its stronger penetration than natural light. When rendering novel view X-ray projections, existing methods mainly based on NeRF suffer from long training time and slow inference speed. In this paper, we propose a 3D Gaussian splatting-based framework, namely X-Gaussian, for X-ray novel view synthesis. Firstly, we redesign a radiative Gaussian point cloud model inspired by the isotropic nature of X-ray imaging. Our model excludes the influence of view direction when learning to predict the radiation intensity of 3D points. Based on this model, we develop a Differentiable Radiative Rasterization (DRR) with CUDA implementation. Secondly, we customize an Angle-pose Cuboid Uniform Initialization (ACUI) strategy that directly uses the parameters of the X-ray scanner to compute the camera information and then uniformly samples point positions within a cuboid enclosing the scanned object. Experiments show that our X-Gaussian outperforms state-of-the-art methods by 6.5 dB while enjoying less than 15% training time and over 73x inference speed. The application on sparse-view CT reconstruction also reveals the practical values of our method. Code is publicly available at https://github.com/caiyuanhao1998/X-Gaussian . A video demo of the training process visualization is at https://www.youtube.com/watch?v=gDVf_Ngeghg .
URLs: https://github.com/caiyuanhao1998/X-Gaussian, https://www.youtube.com/watch?v=gDVf_Ngeghg
Authors: Xinjie Zhang, Xingtong Ge, Tongda Xu, Dailan He, Yan Wang, Hongwei Qin, Guo Lu, Jing Geng, Jun Zhang
Abstract: Implicit neural representations (INRs) recently achieved great success in image representation and compression, offering high visual quality and fast rendering speeds with 10-1000 FPS, assuming sufficient GPU resources are available. However, this requirement often hinders their use on low-end devices with limited memory. In response, we propose a groundbreaking paradigm of image representation and compression by 2D Gaussian Splatting, named GaussianImage. We first introduce 2D Gaussian to represent the image, where each Gaussian has 8 parameters including position, covariance and color. Subsequently, we unveil a novel rendering algorithm based on accumulated summation. Remarkably, our method with a minimum of 3$\times$ lower GPU memory usage and 5$\times$ faster fitting time not only rivals INRs (e.g., WIRE, I-NGP) in representation performance, but also delivers a faster rendering speed of 1500-2000 FPS regardless of parameter size. Furthermore, we integrate existing vector quantization technique to build an image codec. Experimental results demonstrate that our codec attains rate-distortion performance comparable to compression-based INRs such as COIN and COIN++, while facilitating decoding speeds of approximately 2000 FPS. Additionally, preliminary proof of concept shows that our codec surpasses COIN and COIN++ in performance when using partial bits-back coding. Code is available at https://github.com/Xinjie-Q/GaussianImage.
Authors: Weixiang Sun, Yixin Liu, Zhiling Yan, Kaidi Xu, Lichao Sun
Abstract: The rapid expansion of AI in healthcare has led to a surge in medical data generation and storage, boosting medical AI development. However, fears of unauthorized use, like training commercial AI models, hinder researchers from sharing their valuable datasets. To encourage data sharing, one promising solution is to introduce imperceptible noise into the data. This method aims to safeguard the data against unauthorized training by inducing degradation in the generalization ability of the trained model. However, they are not effective and efficient when applied to medical data, mainly due to the ignorance of the sparse nature of medical images. To address this problem, we propose the Sparsity-Aware Local Masking (SALM) method, a novel approach that selectively perturbs significant pixel regions rather than the entire image as previously. This simple yet effective approach, by focusing on local areas, significantly narrows down the search space for disturbances and fully leverages the characteristics of sparsity. Our extensive experiments across various datasets and model architectures demonstrate that SALM effectively prevents unauthorized training of different models and outperforms previous SoTA data protection methods.
Authors: Allen Z. Ren, Jaden Clark, Anushri Dixit, Masha Itkina, Anirudha Majumdar, Dorsa Sadigh
Abstract: We consider the problem of Embodied Question Answering (EQA), which refers to settings where an embodied agent such as a robot needs to actively explore an environment to gather information until it is confident about the answer to a question. In this work, we leverage the strong semantic reasoning capabilities of large vision-language models (VLMs) to efficiently explore and answer such questions. However, there are two main challenges when using VLMs in EQA: they do not have an internal memory for mapping the scene to be able to plan how to explore over time, and their confidence can be miscalibrated and can cause the robot to prematurely stop exploration or over-explore. We propose a method that first builds a semantic map of the scene based on depth information and via visual prompting of a VLM - leveraging its vast knowledge of relevant regions of the scene for exploration. Next, we use conformal prediction to calibrate the VLM's question answering confidence, allowing the robot to know when to stop exploration - leading to a more calibrated and efficient exploration strategy. To test our framework in simulation, we also contribute a new EQA dataset with diverse, realistic human-robot scenarios and scenes built upon the Habitat-Matterport 3D Research Dataset (HM3D). Both simulated and real robot experiments show our proposed approach improves the performance and efficiency over baselines that do no leverage VLM for exploration or do not calibrate its confidence. Webpage with experiment videos and code: https://explore-eqa.github.io/
Authors: Wen Sheng, Zhong Zheng, Jiajun Liu, Han Lu, Hanyuan Zhang, Zhengyong Jiang, Zhihong Zhang, Daoping Zhu
Abstract: Background: Liver tumors are abnormal growths in the liver that can be either benign or malignant, with liver cancer being a significant health concern worldwide. However, there is no dataset for plain scan segmentation of liver tumors, nor any related algorithms. To fill this gap, we propose Plain Scan Liver Tumors(PSLT) and YNetr. Methods: A collection of 40 liver tumor plain scan segmentation datasets was assembled and annotated. Concurrently, we utilized Dice coefficient as the metric for assessing the segmentation outcomes produced by YNetr, having advantage of capturing different frequency information. Results: The YNetr model achieved a Dice coefficient of 62.63% on the PSLT dataset, surpassing the other publicly available model by an accuracy margin of 1.22%. Comparative evaluations were conducted against a range of models including UNet 3+, XNet, UNetr, Swin UNetr, Trans-BTS, COTr, nnUNetv2 (2D), nnUNetv2 (3D fullres), MedNext (2D) and MedNext(3D fullres). Conclusions: We not only proposed a dataset named PSLT(Plain Scan Liver Tumors), but also explored a structure called YNetr that utilizes wavelet transform to extract different frequency information, which having the SOTA in PSLT by experiments.
Authors: Neil Dizon, Jyrki Jauhiainen, Tuomo Valkonen
Abstract: Online optimisation facilitates the solution of dynamic inverse problems, such as image stabilisation, fluid flow monitoring, and dynamic medical imaging. In this paper, we improve upon previous work on predictive online primal-dual methods on two fronts. Firstly, we provide a more concise analysis that symmetrises previously unsymmetric regret bounds, and relaxes previous restrictive conditions on the dual predictor. Secondly, based on the latter, we develop several improved dual predictors. We numerically demonstrate their efficacy in image stabilisation and dynamic positron emission tomography.
Authors: Jayroop Ramesh, Nicola K Dinsdale, the INTERGROWTH-21st Consortium, Pak-Hei Yeung, Ana IL Namburete
Abstract: Accurately localizing two-dimensional (2D) ultrasound (US) fetal brain images in the 3D brain, using minimal computational resources, is an important task for automated US analysis of fetal growth and development. We propose an uncertainty-aware deep learning model for automated 3D plane localization in 2D fetal brain images. Specifically, a multi-head network is trained to jointly regress 3D plane pose from 2D images in terms of different geometric transformations. The model explicitly learns to predict uncertainty to allocate higher weight to inputs with low variances across different transformations to improve performance. Our proposed method, QAERTS, demonstrates superior pose estimation accuracy than the state-of-the-art and most of the uncertainty-based approaches, leading to 9% improvement on plane angle (PA) for localization accuracy, and 8% on normalized cross-correlation (NCC) for sampled image quality. QAERTS also demonstrates efficiency, containing 5$\times$ fewer parameters than ensemble-based approach, making it advantageous in resource-constrained settings. In addition, QAERTS proves to be more robust to noise effects observed in freehand US scanning by leveraging rotational discontinuities and explicit output uncertainties.
Authors: Andy Zou, Long Phan, Justin Wang, Derek Duenas, Maxwell Lin, Maksym Andriushchenko, Rowan Wang, Zico Kolter, Matt Fredrikson, Dan Hendrycks
Abstract: AI systems can take harmful actions and are highly vulnerable to adversarial attacks. We present an approach, inspired by recent advances in representation engineering, that interrupts the models as they respond with harmful outputs with "circuit breakers." Existing techniques aimed at improving alignment, such as refusal training, are often bypassed. Techniques such as adversarial training try to plug these holes by countering specific attacks. As an alternative to refusal training and adversarial training, circuit-breaking directly controls the representations that are responsible for harmful outputs in the first place. Our technique can be applied to both text-only and multimodal language models to prevent the generation of harmful outputs without sacrificing utility -- even in the presence of powerful unseen attacks. Notably, while adversarial robustness in standalone image recognition remains an open challenge, circuit breakers allow the larger multimodal system to reliably withstand image "hijacks" that aim to produce harmful content. Finally, we extend our approach to AI agents, demonstrating considerable reductions in the rate of harmful actions when they are under attack. Our approach represents a significant step forward in the development of reliable safeguards to harmful behavior and adversarial attacks.
Authors: Stefan P. Hein, Manuel Schultheiss, Andrei Gafita, Raphael Zaum, Farid Yagubbayli, Robert Tauber, Isabel Rauscher, Matthias Eiber, Franz Pfeiffer, Wolfgang A. Weber
Abstract: Assessing tumor response to systemic therapies is one of the main applications of PET/CT. Routinely, only a small subset of index lesions out of multiple lesions is analyzed. However, this operator dependent selection may bias the results due to possible significant inter-metastatic heterogeneity of response to therapy. Automated, AI based approaches for lesion tracking hold promise in enabling the analysis of many more lesions and thus providing a better assessment of tumor response. This work introduces a Siamese CNN approach for lesion tracking between PET/CT scans. Our approach is applied on the laborious task of tracking a high number of bone lesions in full-body baseline and follow-up [68Ga]Ga- or [18F]F-PSMA PET/CT scans after two cycles of [177Lu]Lu-PSMA therapy of metastatic castration resistant prostate cancer patients. Data preparation includes lesion segmentation and affine registration. Our algorithm extracts suitable lesion patches and forwards them into a Siamese CNN trained to classify the lesion patch pairs as corresponding or non-corresponding lesions. Experiments have been performed with different input patch types and a Siamese network in 2D and 3D. The CNN model successfully learned to classify lesion assignments, reaching a lesion tracking accuracy of 83 % in its best configuration with an AUC = 0.91. For remaining lesions the pipeline accomplished a re-identification rate of 89 %. We proved that a CNN may facilitate the tracking of multiple lesions in PSMA PET/CT scans. Future clinical studies are necessary if this improves the prediction of the outcome of therapies.
Authors: Jinuk Kim, Marwa El Halabi, Mingi Ji, Hyun Oh Song
Abstract: Recent works show that reducing the number of layers in a convolutional neural network can enhance efficiency while maintaining the performance of the network. Existing depth compression methods remove redundant non-linear activation functions and merge the consecutive convolution layers into a single layer. However, these methods suffer from a critical drawback; the kernel size of the merged layers becomes larger, significantly undermining the latency reduction gained from reducing the depth of the network. We show that this problem can be addressed by jointly pruning convolution layers and activation functions. To this end, we propose LayerMerge, a novel depth compression method that selects which activation layers and convolution layers to remove, to achieve a desired inference speed-up while minimizing performance loss. Since the corresponding selection problem involves an exponential search space, we formulate a novel surrogate optimization problem and efficiently solve it via dynamic programming. Empirical results demonstrate that our method consistently outperforms existing depth compression and layer pruning methods on various network architectures, both on image classification and generation tasks. We release the code at https://github.com/snu-mllab/LayerMerge.
Authors: Long Bai, Tong Chen, Qiaozhi Tan, Wan Jun Nah, Yanheng Li, Zhicheng He, Sishen Yuan, Zhen Chen, Jinlin Wu, Mobarakol Islam, Zhen Li, Hongbin Liu, Hongliang Ren
Abstract: Wireless Capsule Endoscopy (WCE) is highly valued for its non-invasive and painless approach, though its effectiveness is compromised by uneven illumination from hardware constraints and complex internal dynamics, leading to overexposed or underexposed images. While researchers have discussed the challenges of low-light enhancement in WCE, the issue of correcting for different exposure levels remains underexplored. To tackle this, we introduce EndoUIC, a WCE unified illumination correction solution using an end-to-end promptable diffusion transformer (DiT) model. In our work, the illumination prompt module shall navigate the model to adapt to different exposure levels and perform targeted image enhancement, in which the Adaptive Prompt Integration (API) and Global Prompt Scanner (GPS) modules shall further boost the concurrent representation learning between the prompt parameters and features. Besides, the U-shaped restoration DiT model shall capture the long-range dependencies and contextual information for unified illumination restoration. Moreover, we present a novel Capsule-endoscopy Exposure Correction (CEC) dataset, including ground-truth and corrupted image pairs annotated by expert photographers. Extensive experiments against a variety of state-of-the-art (SOTA) methods on four datasets showcase the effectiveness of our proposed method and components in WCE illumination restoration, and the additional downstream experiments further demonstrate its utility for clinical diagnosis and surgical assistance.
Authors: Xingjian Hu, Baole Wei, Liangcai Gao, Jun Wang
Abstract: Text line detection is a key task in historical document analysis facing many challenges of arbitrary-shaped text lines, dense texts, and text lines with high aspect ratios, etc. In this paper, we propose a general framework for historical document text detection (SegHist), enabling existing segmentation-based text detection methods to effectively address the challenges, especially text lines with high aspect ratios. Integrating the SegHist framework with the commonly used method DB++, we develop DB-SegHist. This approach achieves SOTA on the CHDAC, MTHv2, and competitive results on HDRC datasets, with a significant improvement of 1.19% on the most challenging CHDAC dataset which features more text lines with high aspect ratios. Moreover, our method attains SOTA on rotated MTHv2 and rotated HDRC, demonstrating its rotational robustness. The code is available at https://github.com/LumionHXJ/SegHist.
Authors: Chen Wang, Kaiyi Ji, Junyi Geng, Zhongqiang Ren, Taimeng Fu, Fan Yang, Yifan Guo, Haonan He, Xiangyu Chen, Zitong Zhan, Qiwei Du, Shaoshu Su, Bowen Li, Yuheng Qiu, Yi Du, Qihang Li, Yifan Yang, Xiao Lin, Zhipeng Zhao
Abstract: Data-driven methods such as reinforcement and imitation learning have achieved remarkable success in robot autonomy. However, their data-centric nature still hinders them from generalizing well to ever-changing environments. Moreover, collecting large datasets for robotic tasks is often impractical and expensive. To overcome these challenges, we introduce a new self-supervised neural-symbolic (NeSy) computational framework, imperative learning (IL), for robot autonomy, leveraging the generalization abilities of symbolic reasoning. The framework of IL consists of three primary components: a neural module, a reasoning engine, and a memory system. We formulate IL as a special bilevel optimization (BLO), which enables reciprocal learning over the three modules. This overcomes the label-intensive obstacles associated with data-driven approaches and takes advantage of symbolic reasoning concerning logical reasoning, physical principles, geometric analysis, etc. We discuss several optimization techniques for IL and verify their effectiveness in five distinct robot autonomy tasks including path planning, rule induction, optimal control, visual odometry, and multi-robot routing. Through various experiments, we show that IL can significantly enhance robot autonomy capabilities and we anticipate that it will catalyze further research across diverse domains.
Authors: Christophe Bonneville, Nathan Bieberdorf, Arun Hegde, Mark Asta, Habib N. Najm, Laurent Capolungo, Cosmin Safta
Abstract: Prolonged contact between a corrosive liquid and metal alloys can cause progressive dealloying. For such liquid-metal dealloying (LMD) process, phase field models have been developed. However, the governing equations often involve coupled non-linear partial differential equations (PDE), which are challenging to solve numerically. In particular, stiffness in the PDEs requires an extremely small time steps (e.g. $10^{-12}$ or smaller). This computational bottleneck is especially problematic when running LMD simulation until a late time horizon is required. This motivates the development of surrogate models capable of leaping forward in time, by skipping several consecutive time steps at-once. In this paper, we propose U-Shaped Adaptive Fourier Neural Operators (U-AFNO), a machine learning (ML) model inspired by recent advances in neural operator learning. U-AFNO employs U-Nets for extracting and reconstructing local features within the physical fields, and passes the latent space through a vision transformer (ViT) implemented in the Fourier space (AFNO). We use U-AFNOs to learn the dynamics mapping the field at a current time step into a later time step. We also identify global quantities of interest (QoI) describing the corrosion process (e.g. the deformation of the liquid-metal interface) and show that our proposed U-AFNO model is able to accurately predict the field dynamics, in-spite of the chaotic nature of LMD. Our model reproduces the key micro-structure statistics and QoIs with a level of accuracy on-par with the high-fidelity numerical solver. We also investigate the opportunity of using hybrid simulations, in which we alternate forward leap in time using the U-AFNO with high-fidelity time stepping. We demonstrate that while advantageous for some surrogate model design choices, our proposed U-AFNO model in fully auto-regressive settings consistently outperforms hybrid schemes.
Authors: Jing Zou, Lanqing Liu, Qi Chen, Shujun Wang, Zhanli Hu, Xiaohan Xing, Jing Qin
Abstract: Multi-modal MRI offers valuable complementary information for diagnosis and treatment; however, its utility is limited by prolonged scanning times. To accelerate the acquisition process, a practical approach is to reconstruct images of the target modality, which requires longer scanning times, from under-sampled k-space data using the fully-sampled reference modality with shorter scanning times as guidance. The primary challenge of this task is comprehensively and efficiently integrating complementary information from different modalities to achieve high-quality reconstruction. Existing methods struggle with this: 1) convolution-based models fail to capture long-range dependencies; 2) transformer-based models, while excelling in global feature modeling, struggle with quadratic computational complexity. To address this, we propose MMR-Mamba, a novel framework that thoroughly and efficiently integrates multi-modal features for MRI reconstruction, leveraging Mamba's capability to capture long-range dependencies with linear computational complexity while exploiting global properties of the Fourier domain. Specifically, we first design a Target modality-guided Cross Mamba (TCM) module in the spatial domain, which maximally restores the target modality information by selectively incorporating relevant information from the reference modality. Then, we introduce a Selective Frequency Fusion (SFF) module to efficiently integrate global information in the Fourier domain and recover high-frequency signals for the reconstruction of structural details. Furthermore, we devise an Adaptive Spatial-Frequency Fusion (ASFF) module, which mutually enhances the spatial and frequency domains by supplementing less informative channels from one domain with corresponding channels from the other.
Authors: Yiqun Lin, Hualiang Wang, Jixiang Chen, Xiaomeng Li
Abstract: Cone-Beam Computed Tomography (CBCT) is an indispensable technique in medical imaging, yet the associated radiation exposure raises concerns in clinical practice. To mitigate these risks, sparse-view reconstruction has emerged as an essential research direction, aiming to reduce the radiation dose by utilizing fewer projections for CT reconstruction. Although implicit neural representations have been introduced for sparse-view CBCT reconstruction, existing methods primarily focus on local 2D features queried from sparse projections, which is insufficient to process the more complicated anatomical structures, such as the chest. To this end, we propose a novel reconstruction framework, namely DIF-Gaussian, which leverages 3D Gaussians to represent the feature distribution in the 3D space, offering additional 3D spatial information to facilitate the estimation of attenuation coefficients. Furthermore, we incorporate test-time optimization during inference to further improve the generalization capability of the model. We evaluate DIF-Gaussian on two public datasets, showing significantly superior reconstruction performance than previous state-of-the-art methods.
Authors: Joan Nwatu, Oana Ignat, Rada Mihalcea
Abstract: Unequal representation across cultures and socioeconomic groups in AI is a significant and challenging problem, often leading to uneven model performance. As a step toward addressing this issue, we formulate translated non-English, geographic, and socioeconomic integrated prompts and evaluate their impact on VL model performance for data from different countries and income groups. Our findings show that geographic and socioeconomic integrated prompts improve VL performance on lower-income data and favor the retrieval of topic appearances commonly found in data from low-income households. From our analyses, we identify and highlight contexts where these strategies yield the most improvements. Our model analysis code is publicly available at https://github.com/Anniejoan/Uplifting-Lower-income-data .
URLs: https://github.com/Anniejoan/Uplifting-Lower-income-data
Authors: Yunpeng Jiang, Paul Weng, Yutong Ban
Abstract: Data augmentation is widely applied and has shown its benefits in different machine learning tasks. However, as recently observed in some downstream tasks, data augmentation may introduce an unfair impact on classifications. While it can improve the performance of some classes, it can actually be detrimental for other classes, which can be problematic in some application domains. In this paper, to counteract this phenomenon, we propose a FAir Classification approach with a Two-player game (FACT). We first formulate the training of a classifier with data augmentation as a fair optimization problem, which can be further written as an adversarial two-player game. Following this formulation, we propose a novel multiplicative weight optimization algorithm, for which we theoretically prove that it can converge to a solution that is fair over classes. Interestingly, our formulation also reveals that this fairness issue over classes is not due to data augmentation only, but is in fact a general phenomenon. Our empirical experiments demonstrate that the performance of our learned classifiers is indeed more fairly distributed over classes in five datasets, with only limited impact on the average accuracy.