Authors: Andrew Seohwan Yu, Mohsen Hariri, Xuecen Zhang, Mingrui Yang, Vipin Chaudhary, Xiaojuan Li
Abstract: Intelligent medical image segmentation methods are rapidly evolving and being increasingly applied, yet they face the challenge of domain transfer, where algorithm performance degrades due to different data distributions between source and target domains. To address this, we introduce a method for zero-shot, single-prompt segmentation of 3D knee MRI by adapting Segment Anything Model 2 (SAM2), a general-purpose segmentation model designed to accept prompts and retain memory across frames of a video. By treating slices from 3D medical volumes as individual video frames, we leverage SAM2's advanced capabilities to generate motion- and spatially-aware predictions. We demonstrate that SAM2 can efficiently perform segmentation tasks in a zero-shot manner with no additional training or fine-tuning, accurately delineating structures in knee MRI scans using only a single prompt. Our experiments on the Osteoarthritis Initiative Zuse Institute Berlin (OAI-ZIB) dataset reveal that SAM2 achieves high accuracy on 3D knee bone segmentation, with a testing Dice similarity coefficient of 0.9643 on tibia. We also present results generated using different SAM2 model sizes, different prompt schemes, as well as comparative results from the SAM1 model deployed on the same dataset. This breakthrough has the potential to revolutionize medical image analysis by providing a scalable, cost-effective solution for automated segmentation, paving the way for broader clinical applications and streamlined workflows.
Authors: Saurabh Farkya, Zachary Alan Daniels, Aswin Raghavan, Gooitzen van der Wal, Michael Isnardi, Michael Piacentino, David Zhang
Abstract: Recent advancements in sensors have led to high resolution and high data throughput at the pixel level. Simultaneously, the adoption of increasingly large (deep) neural networks (NNs) has lead to significant progress in computer vision. Currently, visual intelligence comes at increasingly high computational complexity, energy, and latency. We study a data-driven system that combines dynamic sensing at the pixel level with computer vision analytics at the video level and propose a feedback control loop to minimize data movement between the sensor front-end and computational back-end without compromising detection and tracking precision. Our contributions are threefold: (1) We introduce anticipatory attention and show that it leads to high precision prediction with sparse activation of pixels; (2) Leveraging the feedback control, we show that the dimensionality of learned feature vectors can be significantly reduced with increased sparsity; and (3) We emulate analog design choices (such as varying RGB or Bayer pixel format and analog noise) and study their impact on the key metrics of the data-driven system. Comparative analysis with traditional pixel and deep learning models shows significant performance enhancements. Our system achieves a 10X reduction in bandwidth and a 15-30X improvement in Energy-Delay Product (EDP) when activating only 30% of pixels, with a minor reduction in object detection and tracking precision. Based on analog emulation, our system can achieve a throughput of 205 megapixels/sec (MP/s) with a power consumption of only 110 mW per MP, i.e., a theoretical improvement of ~30X in EDP.
Authors: James Baker
Abstract: Textual Inversion remains a popular method for personalizing diffusion models, in order to teach models new subjects and styles. We note that textual inversion has been underexplored using alternatives to the UNet, and experiment with textual inversion with a vision transformer. We also seek to optimize textual inversion using a strategy that does not require explicit use of the UNet and its idiosyncratic layers, so we add bonus tokens and enforce orthogonality. We find the use of the bonus token improves adherence to the source images and the use of the vision transformer improves adherence to the prompt. Code is available at https://github.com/jamesBaker361/tex_inv_plus.
Authors: Boshra Khalili, Andrew W. Smyth
Abstract: Object detection as part of computer vision can be crucial for traffic management, emergency response, autonomous vehicles, and smart cities. Despite significant advances in object detection, detecting small objects in images captured by distant cameras remains challenging due to their size, distance from the camera, varied shapes, and cluttered backgrounds. To address these challenges, we propose Small Object Detection YOLOv8 (SOD-YOLOv8), a novel model specifically designed for scenarios involving numerous small objects. Inspired by Efficient Generalized Feature Pyramid Networks (GFPN), we enhance multi-path fusion within YOLOv8 to integrate features across different levels, preserving details from shallower layers and improving small object detection accuracy. Also, A fourth detection layer is added to leverage high-resolution spatial information effectively. The Efficient Multi-Scale Attention Module (EMA) in the C2f-EMA module enhances feature extraction by redistributing weights and prioritizing relevant features. We introduce Powerful-IoU (PIoU) as a replacement for CIoU, focusing on moderate-quality anchor boxes and adding a penalty based on differences between predicted and ground truth bounding box corners. This approach simplifies calculations, speeds up convergence, and enhances detection accuracy. SOD-YOLOv8 significantly improves small object detection, surpassing widely used models in various metrics, without substantially increasing computational cost or latency compared to YOLOv8s. Specifically, it increases recall from 40.1\% to 43.9\%, precision from 51.2\% to 53.9\%, $\text{mAP}_{0.5}$ from 40.6\% to 45.1\%, and $\text{mAP}_{0.5:0.95}$ from 24\% to 26.6\%. In dynamic real-world traffic scenes, SOD-YOLOv8 demonstrated notable improvements in diverse conditions, proving its reliability and effectiveness in detecting small objects even in challenging environments.
Authors: Piraveen Sivakumar, Paul Janson, Jathushan Rajasegaran, Thanuja Ambegoda
Abstract: In this paper, we address the challenge of generating novel views of real-world objects with limited multi-view images through our proposed approach, FewShotNeRF. Our method utilizes meta-learning to acquire optimal initialization, facilitating rapid adaptation of a Neural Radiance Field (NeRF) to specific scenes. The focus of our meta-learning process is on capturing shared geometry and textures within a category, embedded in the weight initialization. This approach expedites the learning process of NeRFs and leverages recent advancements in positional encodings to reduce the time required for fitting a NeRF to a scene, thereby accelerating the inner loop optimization of meta-learning. Notably, our method enables meta-learning on a large number of 3D scenes to establish a robust 3D prior for various categories. Through extensive evaluations on the Common Objects in 3D open source dataset, we empirically demonstrate the efficacy and potential of meta-learning in generating high-quality novel views of objects.
Authors: Yifan Feng, Jiangang Huang, Shaoyi Du, Shihui Ying, Jun-Hai Yong, Yipeng Li, Guiguang Ding, Rongrong Ji, Yue Gao
Abstract: We introduce Hyper-YOLO, a new object detection method that integrates hypergraph computations to capture the complex high-order correlations among visual features. Traditional YOLO models, while powerful, have limitations in their neck designs that restrict the integration of cross-level features and the exploitation of high-order feature interrelationships. To address these challenges, we propose the Hypergraph Computation Empowered Semantic Collecting and Scattering (HGC-SCS) framework, which transposes visual feature maps into a semantic space and constructs a hypergraph for high-order message propagation. This enables the model to acquire both semantic and structural information, advancing beyond conventional feature-focused learning. Hyper-YOLO incorporates the proposed Mixed Aggregation Network (MANet) in its backbone for enhanced feature extraction and introduces the Hypergraph-Based Cross-Level and Cross-Position Representation Network (HyperC2Net) in its neck. HyperC2Net operates across five scales and breaks free from traditional grid structures, allowing for sophisticated high-order interactions across levels and positions. This synergy of components positions Hyper-YOLO as a state-of-the-art architecture in various scale models, as evidenced by its superior performance on the COCO dataset. Specifically, Hyper-YOLO-N significantly outperforms the advanced YOLOv8-N and YOLOv9-T with 12\% $\text{AP}^{val}$ and 9\% $\text{AP}^{val}$ improvements. The source codes are at ttps://github.com/iMoonLab/Hyper-YOLO.
Authors: Haider Al-Tahan, Quentin Garrido, Randall Balestriero, Diane Bouchacourt, Caner Hazirbas, Mark Ibrahim
Abstract: Significant research efforts have been made to scale and improve vision-language model (VLM) training approaches. Yet, with an ever-growing number of benchmarks, researchers are tasked with the heavy burden of implementing each protocol, bearing a non-trivial computational cost, and making sense of how all these benchmarks translate into meaningful axes of progress. To facilitate a systematic evaluation of VLM progress, we introduce UniBench: a unified implementation of 50+ VLM benchmarks spanning a comprehensive range of carefully categorized capabilities from object recognition to spatial awareness, counting, and much more. We showcase the utility of UniBench for measuring progress by evaluating nearly 60 publicly available vision-language models, trained on scales of up to 12.8B samples. We find that while scaling training data or model size can boost many vision-language model capabilities, scaling offers little benefit for reasoning or relations. Surprisingly, we also discover today's best VLMs struggle on simple digit recognition and counting tasks, e.g. MNIST, which much simpler networks can solve. Where scale falls short, we find that more precise interventions, such as data quality or tailored-learning objectives offer more promise. For practitioners, we also offer guidance on selecting a suitable VLM for a given application. Finally, we release an easy-to-run UniBench code-base with the full set of 50+ benchmarks and comparisons across 59 models as well as a distilled, representative set of benchmarks that runs in 5 minutes on a single GPU.
Authors: Yingfan Ma, Xiaoyuan Luo, Mingzhi Yuan, Xinrong Chen, Manning Wang
Abstract: Multiple instance learning (MIL) problem is currently solved from either bag-classification or instance-classification perspective, both of which ignore important information contained in some instances and result in limited performance. For example, existing methods often face difficulty in learning hard positive instances. In this paper, we formulate MIL as a semi-supervised instance classification problem, so that all the labeled and unlabeled instances can be fully utilized to train a better classifier. The difficulty in this formulation is that all the labeled instances are negative in MIL, and traditional self-training techniques used in semi-supervised learning tend to degenerate in generating pseudo labels for the unlabeled instances in this scenario. To resolve this problem, we propose a weakly-supervised self-training method, in which we utilize the positive bag labels to construct a global constraint and a local constraint on the pseudo labels to prevent them from degenerating and force the classifier to learn hard positive instances. It is worth noting that easy positive instances are instances are far from the decision boundary in the classification process, while hard positive instances are those close to the decision boundary. Through iterative optimization, the pseudo labels can gradually approach the true labels. Extensive experiments on two MNIST synthetic datasets, five traditional MIL benchmark datasets and two histopathology whole slide image datasets show that our method achieved new SOTA performance on all of them. The code will be publicly available.
Authors: Bingbing Dan, Meihui Li, Tao Tang, Jing Zhang
Abstract: Infrared small target sequences exhibit strong similarities between frames and contain rich contextual information, which motivates us to achieve sequential infrared small target segmentation with minimal data. Inspired by the success of large segmentation models led by Segment Anything Model (SAM) across various downstream tasks, we propose a one-shot and training-free method that perfectly adapts SAM's zero-shot generalization capabilities to sequential infrared small target segmentation. Given one annotated frame as a reference, our method can accurately segment small targets in other frames of the sequence. Specifically, we first obtain a confidence map through local feature matching between reference image and test image. Then, the highest point in the confidence map is as a prompt, and we design the Point Prompt-Centric Focusing (PPCF) module to address the over-segmentation of small targets with blurry boundaries. Subsequently, to prevent miss and false detections, we introduce the Triple-Level Ensemble (TLE) module that ensembles the masks obtained at different levels from the first two steps to produce the final mask. Experiments demonstrate that our method requires only one shot to achieve comparable performance to state-of-the-art methods based on traditional many-shot supervision and even superior performance in a few-shot setting. Moreover, ablation studies confirm the robustness of our approach to variations in one-shot samples, changes in scenes, and the presence of multiple targets.
Authors: Lingbei Meng, Bi'an Du, Wei Hu
Abstract: Sparse-view 3D reconstruction stands as a formidable challenge in computer vision, aiming to build complete three-dimensional models from a limited array of viewing perspectives. This task confronts several difficulties: 1) the limited number of input images that lack consistent information; 2) dependence on the quality of input images; and 3) the substantial size of model parameters. To address these challenges, we propose a self-augmented coarse-to-fine Gaussian splatting paradigm, enhanced with a structure-aware mask, for sparse-view 3D reconstruction. In particular, our method initially employs a coarse Gaussian model to obtain a basic 3D representation from sparse-view inputs. Subsequently, we develop a fine Gaussian network to enhance consistent and detailed representation of the output with both 3D geometry augmentation and perceptual view augmentation. During training, we design a structure-aware masking strategy to further improve the model's robustness against sparse inputs and noise.Experimental results on the MipNeRF360 and OmniObject3D datasets demonstrate that the proposed method achieves state-of-the-art performances for sparse input views in both perceptual quality and efficiency.
Authors: Jiabo Ye, Haiyang Xu, Haowei Liu, Anwen Hu, Ming Yan, Qi Qian, Ji Zhang, Fei Huang, Jingren Zhou
Abstract: Multi-modal Large Language Models (MLLMs) have demonstrated remarkable capabilities in executing instructions for a variety of single-image tasks. Despite this progress, significant challenges remain in modeling long image sequences. In this work, we introduce the versatile multi-modal large language model, mPLUG-Owl3, which enhances the capability for long image-sequence understanding in scenarios that incorporate retrieved image-text knowledge, interleaved image-text, and lengthy videos. Specifically, we propose novel hyper attention blocks to efficiently integrate vision and language into a common language-guided semantic space, thereby facilitating the processing of extended multi-image scenarios. Extensive experimental results suggest that mPLUG-Owl3 achieves state-of-the-art performance among models with a similar size on single-image, multi-image, and video benchmarks. Moreover, we propose a challenging long visual sequence evaluation named Distractor Resistance to assess the ability of models to maintain focus amidst distractions. Finally, with the proposed architecture, mPLUG-Owl3 demonstrates outstanding performance on ultra-long visual sequence inputs. We hope that mPLUG-Owl3 can contribute to the development of more efficient and powerful multimodal large language models.
Authors: Parisa Farmanifard, Arun Ross
Abstract: This study utilizes the advanced capabilities of the GPT-4 multimodal Large Language Model (LLM) to explore its potential in iris recognition - a field less common and more specialized than face recognition. By focusing on this niche yet crucial area, we investigate how well AI tools like ChatGPT can understand and analyze iris images. Through a series of meticulously designed experiments employing a zero-shot learning approach, the capabilities of ChatGPT-4 was assessed across various challenging conditions including diverse datasets, presentation attacks, occlusions such as glasses, and other real-world variations. The findings convey ChatGPT-4's remarkable adaptability and precision, revealing its proficiency in identifying distinctive iris features, while also detecting subtle effects like makeup on iris recognition. A comparative analysis with Gemini Advanced - Google's AI model - highlighted ChatGPT-4's better performance and user experience in complex iris analysis tasks. This research not only validates the use of LLMs for specialized biometric applications but also emphasizes the importance of nuanced query framing and interaction design in extracting significant insights from biometric data. Our findings suggest a promising path for future research and the development of more adaptable, efficient, robust and interactive biometric security solutions.
Authors: Jingcai Guo, Zhijie Rao, Zhi Chen, Song Guo, Jingren Zhou, Dacheng Tao
Abstract: Zero-shot image recognition (ZSIR) aims at empowering models to recognize and reason in unseen domains via learning generalized knowledge from limited data in the seen domain. The gist for ZSIR is to execute element-wise representation and reasoning from the input visual space to the target semantic space, which is a bottom-up modeling paradigm inspired by the process by which humans observe the world, i.e., capturing new concepts by learning and combining the basic components or shared characteristics. In recent years, element-wise learning techniques have seen significant progress in ZSIR as well as widespread application. However, to the best of our knowledge, there remains a lack of a systematic overview of this topic. To enrich the literature and provide a sound basis for its future development, this paper presents a broad review of recent advances in element-wise ZSIR. Concretely, we first attempt to integrate the three basic ZSIR tasks of object recognition, compositional recognition, and foundation model-based open-world recognition into a unified element-wise perspective and provide a detailed taxonomy and analysis of the main research approaches. Then, we collect and summarize some key information and benchmarks, such as detailed technical implementations and common datasets. Finally, we sketch out the wide range of its related applications, discuss vital challenges, and suggest potential future directions.
Authors: Mengcheng Lan, Chaofeng Chen, Yiping Ke, Xinjiang Wang, Litong Feng, Wayne Zhang
Abstract: Open-vocabulary semantic segmentation requires models to effectively integrate visual representations with open-vocabulary semantic labels. While Contrastive Language-Image Pre-training (CLIP) models shine in recognizing visual concepts from text, they often struggle with segment coherence due to their limited localization ability. In contrast, Vision Foundation Models (VFMs) excel at acquiring spatially consistent local visual representations, yet they fall short in semantic understanding. This paper introduces ProxyCLIP, an innovative framework designed to harmonize the strengths of both CLIP and VFMs, facilitating enhanced open-vocabulary semantic segmentation. ProxyCLIP leverages the spatial feature correspondence from VFMs as a form of proxy attention to augment CLIP, thereby inheriting the VFMs' robust local consistency and maintaining CLIP's exceptional zero-shot transfer capacity. We propose an adaptive normalization and masking strategy to get the proxy attention from VFMs, allowing for adaptation across different VFMs. Remarkably, as a training-free approach, ProxyCLIP significantly improves the average mean Intersection over Union (mIoU) across eight benchmarks from 40.3 to 44.4, showcasing its exceptional efficacy in bridging the gap between spatial precision and semantic richness for the open-vocabulary segmentation task.
Authors: Toshiyuki Oshima, Kentaro Takagi, Kouta Nakata
Abstract: Recently, representation learning with contrastive learning algorithms has been successfully applied to challenging unlabeled datasets. However, these methods are unable to distinguish important features from unimportant ones under simply unsupervised settings, and definitions of importance vary according to the type of downstream task or analysis goal, such as the identification of objects or backgrounds. In this paper, we focus on unsupervised image clustering as the downstream task and propose a representation learning method that enhances features critical to the clustering task. We extend a clustering-friendly contrastive learning method and incorporate a contrastive analysis approach, which utilizes a reference dataset to separate important features from unimportant ones, into the design of loss functions. Conducting an experimental evaluation of image clustering for three datasets with characteristic backgrounds, we show that for all datasets, our method achieves higher clustering scores compared with conventional contrastive analysis and deep clustering methods.
Authors: Haochen Zhao, Hui Meng, Deqian Yang, Xiaozheng Xie, Xiaoze Wu, Qingfeng Li, Jianwei Niu
Abstract: Semi-supervised multi-organ medical image segmentation aids physicians in improving disease diagnosis and treatment planning and reduces the time and effort required for organ annotation.Existing state-of-the-art methods train the labeled data with ground truths and train the unlabeled data with pseudo-labels. However, the two training flows are separate, which does not reflect the interrelationship between labeled and unlabeled data.To address this issue, we propose a semi-supervised multi-organ segmentation method called GuidedNet, which leverages the knowledge from labeled data to guide the training of unlabeled data. The primary goals of this study are to improve the quality of pseudo-labels for unlabeled data and to enhance the network's learning capability for both small and complex organs.A key concept is that voxel features from labeled and unlabeled data that are close to each other in the feature space are more likely to belong to the same class.On this basis, a 3D Consistent Gaussian Mixture Model (3D-CGMM) is designed to leverage the feature distributions from labeled data to rectify the generated pseudo-labels.Furthermore, we introduce a Knowledge Transfer Cross Pseudo Supervision (KT-CPS) strategy, which leverages the prior knowledge obtained from the labeled data to guide the training of the unlabeled data, thereby improving the segmentation accuracy for both small and complex organs. Extensive experiments on two public datasets, FLARE22 and AMOS, demonstrated that GuidedNet is capable of achieving state-of-the-art performance.
Authors: Jaehyuk Heo, Pilsung Kang
Abstract: Active learning (AL) aims to enhance model performance by selectively collecting highly informative data, thereby minimizing annotation costs. However, in practical scenarios, unlabeled data may contain out-of-distribution (OOD) samples, leading to wasted annotation costs if data is incorrectly selected. Recent research has explored methods to apply AL to open-set data, but these methods often require or incur unavoidable cost losses to minimize them. To address these challenges, we propose a novel selection strategy, CLIPN for AL (CLIPNAL), which minimizes cost losses without requiring OOD samples. CLIPNAL sequentially evaluates the purity and informativeness of data. First, it utilizes a pre-trained vision-language model to detect and exclude OOD data by leveraging linguistic and visual information of in-distribution (ID) data without additional training. Second, it selects highly informative data from the remaining ID data, and then the selected samples are annotated by human experts. Experimental results on datasets with various open-set conditions demonstrate that CLIPNAL achieves the lowest cost loss and highest performance across all scenarios. Code is available at https://github.com/DSBA-Lab/OpenAL.
Authors: Ragib Amin Nihal, Benjamin Yen, Katsutoshi Itoyama, Kazuhiro Nakadai
Abstract: Unmanned aerial vehicles (UAVs) have revolutionized search and rescue (SAR) operations, but the lack of specialized human detection datasets for training machine learning models poses a significant challenge.To address this gap, this paper introduces the Combination to Application (C2A) dataset, synthesized by overlaying human poses onto UAV-captured disaster scenes. Through extensive experimentation with state-of-the-art detection models, we demonstrate that models fine-tuned on the C2A dataset exhibit substantial performance improvements compared to those pre-trained on generic aerial datasets. Furthermore, we highlight the importance of combining the C2A dataset with general human datasets to achieve optimal performance and generalization across various scenarios. This points out the crucial need for a tailored dataset to enhance the effectiveness of SAR operations. Our contributions also include developing dataset creation pipeline and integrating diverse human poses and disaster scenes information to assess the severity of disaster scenarios. Our findings advocate for future developments, to ensure that SAR operations benefit from the most realistic and effective AI-assisted interventions possible.
Authors: Palak Handa, Amirreza Mahbod, Florian Schwarzhans, Ramona Woitek, Nidhi Goel, Deepti Chhabra, Shreshtha Jha, Manas Dhir, Deepak Gunjan, Jagadeesh Kakarla, Balasubramanian Raman
Abstract: We present the Capsule Vision 2024 Challenge: Multi-Class Abnormality Classification for Video Capsule Endoscopy. It is being virtually organized by the Research Center for Medical Image Analysis and Artificial Intelligence (MIAAI), Department of Medicine, Danube Private University, Krems, Austria and Medical Imaging and Signal Analysis Hub (MISAHUB) in collaboration with the 9th International Conference on Computer Vision & Image Processing (CVIP 2024) being organized by the Indian Institute of Information Technology, Design and Manufacturing (IIITDM) Kancheepuram, Chennai, India. This document describes the overview of the challenge, its registration and rules, submission format, and the description of the utilized datasets.
Authors: Yizhang Jin, Jian Li, Jiangning Zhang, Jianlong Hu, Zhenye Gan, Xin Tan, Yong Liu, Yabiao Wang, Chengjie Wang, Lizhuang Ma
Abstract: Visual Spatial Description (VSD) aims to generate texts that describe the spatial relationships between objects within images. Traditional visual spatial relationship classification (VSRC) methods typically output the spatial relationship between two objects in an image, often neglecting world knowledge and lacking general language capabilities. In this paper, we propose a Large Language-and-Vision Assistant for Visual Spatial Description, named LLaVA-VSD, which is designed for the classification, description, and open-ended description of visual spatial relationships. Specifically, the model first constructs a VSD instruction-following dataset using given figure-caption pairs for the three tasks. It then employs LoRA to fine-tune a Large Language and Vision Assistant for VSD, which has 13 billion parameters and supports high-resolution images. Finally, a large language model (Qwen-2) is used to refine the generated sentences, enhancing their diversity and accuracy. LLaVA-VSD demonstrates excellent multimodal conversational capabilities and can follow open-ended instructions to assist with inquiries about object relationships in images.
Authors: Long Bai, Guankun Wang, Mobarakol Islam, Lalithkumar Seenivasan, An Wang, Hongliang Ren
Abstract: Medical visual question answering (VQA) bridges the gap between visual information and clinical decision-making, enabling doctors to extract understanding from clinical images and videos. In particular, surgical VQA can enhance the interpretation of surgical data, aiding in accurate diagnoses, effective education, and clinical interventions. However, the inability of VQA models to visually indicate the regions of interest corresponding to the given questions results in incomplete comprehension of the surgical scene. To tackle this, we propose the surgical visual question localized-answering (VQLA) for precise and context-aware responses to specific queries regarding surgical images. Furthermore, to address the strong demand for safety in surgical scenarios and potential corruptions in image acquisition and transmission, we propose a novel approach called Calibrated Co-Attention Gated Vision-Language (C$^2$G-ViL) embedding to integrate and align multimodal information effectively. Additionally, we leverage the adversarial sample-based contrastive learning strategy to boost our performance and robustness. We also extend our EndoVis-18-VQLA and EndoVis-17-VQLA datasets to broaden the scope and application of our data. Extensive experiments on the aforementioned datasets demonstrate the remarkable performance and robustness of our solution. Our solution can effectively combat real-world image corruption. Thus, our proposed approach can serve as an effective tool for assisting surgical education, patient care, and enhancing surgical outcomes.
Authors: Dahyun Kang, Minsu Cho
Abstract: We present lazy visual grounding, a two-stage approach of unsupervised object mask discovery followed by object grounding, for open-vocabulary semantic segmentation. Plenty of the previous art casts this task as pixel-to-text classification without object-level comprehension, leveraging the image-to-text classification capability of pretrained vision-and-language models. We argue that visual objects are distinguishable without the prior text information as segmentation is essentially a vision task. Lazy visual grounding first discovers object masks covering an image with iterative Normalized cuts and then later assigns text on the discovered objects in a late interaction manner. Our model requires no additional training yet shows great performance on five public datasets: Pascal VOC, Pascal Context, COCO-object, COCO-stuff, and ADE 20K. Especially, the visually appealing segmentation results demonstrate the model capability to localize objects precisely. Paper homepage: https://cvlab.postech.ac.kr/research/lazygrounding
Authors: Jihoon Lee, Yunhong Min, Hwidong Kim, Sangtae Ahn
Abstract: In recent years, there has been a significant focus on research related to text-guided image inpainting. However, the task remains challenging due to several constraints, such as ensuring alignment between the image and the text, and maintaining consistency in distribution between corrupted and uncorrupted regions. In this paper, thus, we propose a dual affine transformation generative adversarial network (DAFT-GAN) to maintain the semantic consistency for text-guided inpainting. DAFT-GAN integrates two affine transformation networks to combine text and image features gradually for each decoding block. Moreover, we minimize information leakage of uncorrupted features for fine-grained image generation by encoding corrupted and uncorrupted regions of the masked image separately. Our proposed model outperforms the existing GAN-based models in both qualitative and quantitative assessments with three benchmark datasets (MS-COCO, CUB, and Oxford) for text-guided image inpainting.
Authors: Hangyu Li, Xiangxiang Chu, Dingyuan Shi
Abstract: The Score Distillation Sampling (SDS), which exploits pretrained text-to-image model diffusion models as priors to 3D model training, has achieved significant success. Currently, the flow-based diffusion model has become a new trend for generations. Yet, adapting SDS to flow-based diffusion models in 3D generation remains unexplored. Our work is aimed to bridge this gap. In this paper, we adapt SDS to rectified flow and re-examine the over-smoothing issue under this novel framework. The issue can be explained that the model learns an average of multiple ODE trajectories. Then we propose DreamCouple, which instead of randomly sampling noise, uses a rectified flow model to find the coupled noise. Its Unique Couple Matching (UCM) loss guides the model to learn different trajectories and thus solves the over-smoothing issue. We apply our method to both NeRF and 3D Gaussian splatting and achieve state-of-the-art performances. We also identify some other interesting open questions such as initialization issues for NeRF and faster training convergence. Our code will be released soon.
Authors: Dongsheng Wang, Jiequan Cui, Miaoge Li, Wang Lin, Bo Chen, Hanwang Zhang
Abstract: As the open community of large language models (LLMs) matures, multimodal LLMs (MLLMs) have promised an elegant bridge between vision and language. However, current research is inherently constrained by challenges such as the need for high-quality instruction pairs and the loss of visual information in image-to-text training objectives. To this end, we propose a Visual Token Complement framework (VTC) that helps MLLMs regain the missing visual features and thus improve response accuracy. Specifically, our VTC integrates text-to-image generation as a guide to identifying the text-irrelevant features, and a visual selector is then developed to generate complementary visual tokens to enrich the original visual input. Moreover, an iterative strategy is further designed to extract more visual information by iteratively using the visual selector without any additional training. Notably, the training pipeline requires no additional image-text pairs, resulting in a desired instruction tuning-free property. Both qualitative and quantitative experiments demonstrate the superiority and efficiency of our VTC.
Authors: Alexander Musiat, Laurenz Reichardt, Michael Schulze, Oliver Wasenm\"uller
Abstract: Automotive radar systems have evolved to provide not only range, azimuth and Doppler velocity, but also elevation data. This additional dimension allows for the representation of 4D radar as a 3D point cloud. As a result, existing deep learning methods for 3D object detection, which were initially developed for LiDAR data, are often applied to these radar point clouds. However, this neglects the special characteristics of 4D radar data, such as the extreme sparsity and the optimal utilization of velocity information. To address these gaps in the state-of-the-art, we present RadarPillars, a pillar-based object detection network. By decomposing radial velocity data, introducing PillarAttention for efficient feature extraction, and studying layer scaling to accommodate radar sparsity, RadarPillars significantly outperform state-of-the-art detection results on the View-of-Delft dataset. Importantly, this comes at a significantly reduced parameter count, surpassing existing methods in terms of efficiency and enabling real-time performance on edge devices.
Authors: Zijian Zhu, Ali Zia, Xuesong Li, Bingbing Dan, Yuebo Ma, Hongfeng Long, Kaili Lu, Enhai Liu, Rujin Zhao
Abstract: Stripe-like space target detection (SSTD) is crucial for space situational awareness. Traditional unsupervised methods often fail in low signal-to-noise ratio and variable stripe-like space targets scenarios, leading to weak generalization. Although fully supervised learning methods improve model generalization, they require extensive pixel-level labels for training. In the SSTD task, manually creating these labels is often inaccurate and labor-intensive. Semi-supervised learning (SSL) methods reduce the need for these labels and enhance model generalizability, but their performance is limited by pseudo-label quality. To address this, we introduce an innovative Collaborative Static-Dynamic Teacher (CSDT) SSL framework, which includes static and dynamic teacher models as well as a student model. This framework employs a customized adaptive pseudo-labeling (APL) strategy, transitioning from initial static teaching to adaptive collaborative teaching, guiding the student model's training. The exponential moving average (EMA) mechanism further enhances this process by feeding new stripe-like knowledge back to the dynamic teacher model through the student model, creating a positive feedback loop that continuously enhances the quality of pseudo-labels. Moreover, we present MSSA-Net, a novel SSTD network featuring a multi-scale dual-path convolution (MDPC) block and a feature map weighted attention (FMWA) block, designed to extract diverse stripe-like features within the CSDT SSL training framework. Extensive experiments verify the state-of-the-art performance of our framework on the AstroStripeSet and various ground-based and space-based real-world datasets.
Authors: Daniel Ortega de Carvalho, Luiz Felipe Teodoro Monteiro, Fernanda Marques Bazilio, Gabriel Toshio Hirokawa Higa, Hemerson Pistori
Abstract: Counting fish larvae is an important, yet demanding and time consuming, task in aquaculture. In order to address this problem, in this work, we evaluate four neural network architectures, including convolutional neural networks and transformers, in different sizes, in the task of fish larvae counting. For the evaluation, we present a new annotated image dataset with less data collection requirements than preceding works, with images of spotted sorubim and dourado larvae. By using image tiling techniques, we achieve a MAPE of 4.46% ($\pm 4.70$) with an extra large real time detection transformer, and 4.71% ($\pm 4.98$) with a medium-sized YOLOv8.
Authors: Zeyu Yang, Nan Song, Wei Li, Xiatian Zhu, Li Zhang, Philip H. S. Torr
Abstract: Existing top-performance autonomous driving systems typically rely on the multi-modal fusion strategy for reliable scene understanding. This design is however fundamentally restricted due to overlooking the modality-specific strengths and finally hampering the model performance. To address this limitation, in this work, we introduce a novel modality interaction strategy that allows individual per-modality representations to be learned and maintained throughout, enabling their unique characteristics to be exploited during the whole perception pipeline. To demonstrate the effectiveness of the proposed strategy, we design DeepInteraction++, a multi-modal interaction framework characterized by a multi-modal representational interaction encoder and a multi-modal predictive interaction decoder. Specifically, the encoder is implemented as a dual-stream Transformer with specialized attention operation for information exchange and integration between separate modality-specific representations. Our multi-modal representational learning incorporates both object-centric, precise sampling-based feature alignment and global dense information spreading, essential for the more challenging planning task. The decoder is designed to iteratively refine the predictions by alternately aggregating information from separate representations in a unified modality-agnostic manner, realizing multi-modal predictive interaction. Extensive experiments demonstrate the superior performance of the proposed framework on both 3D object detection and end-to-end autonomous driving tasks. Our code is available at https://github.com/fudan-zvg/DeepInteraction.
Authors: Rishubh Parihar, Sachidanand VS, Sabariswaran Mani, Tejan Karmali, R. Venkatesh Babu
Abstract: Recently, we have seen a surge of personalization methods for text-to-image (T2I) diffusion models to learn a concept using a few images. Existing approaches, when used for face personalization, suffer to achieve convincing inversion with identity preservation and rely on semantic text-based editing of the generated face. However, a more fine-grained control is desired for facial attribute editing, which is challenging to achieve solely with text prompts. In contrast, StyleGAN models learn a rich face prior and enable smooth control towards fine-grained attribute editing by latent manipulation. This work uses the disentangled $\mathcal{W+}$ space of StyleGANs to condition the T2I model. This approach allows us to precisely manipulate facial attributes, such as smoothly introducing a smile, while preserving the existing coarse text-based control inherent in T2I models. To enable conditioning of the T2I model on the $\mathcal{W+}$ space, we train a latent mapper to translate latent codes from $\mathcal{W+}$ to the token embedding space of the T2I model. The proposed approach excels in the precise inversion of face images with attribute preservation and facilitates continuous control for fine-grained attribute editing. Furthermore, our approach can be readily extended to generate compositions involving multiple individuals. We perform extensive experiments to validate our method for face personalization and fine-grained attribute editing.
Authors: Mert Bulent Sariyildiz, Philippe Weinzaepfel, Thomas Lucas, Diane Larlus, Yannis Kalantidis
Abstract: Pretrained models have become a commodity and offer strong results on a broad range of tasks. In this work, we focus on classification and seek to learn a unique encoder able to take from several complementary pretrained models. We aim at even stronger generalization across a variety of classification tasks. We propose to learn such an encoder via multi-teacher distillation. We first thoroughly analyse standard distillation when driven by multiple strong teachers with complementary strengths. Guided by this analysis, we gradually propose improvements to the basic distillation setup. Among those, we enrich the architecture of the encoder with a ladder of expendable projectors, which increases the impact of intermediate features during distillation, and we introduce teacher dropping, a regularization mechanism that better balances the teachers' influence. Our final distillation strategy leads to student models of the same capacity as any of the teachers, while retaining or improving upon the performance of the best teacher for each task. Project page and code: https://europe.naverlabs.com/unic
Authors: Huilin Tian, Jingke Meng, Wei-Shi Zheng, Yuan-Ming Li, Junkai Yan, Yunong Zhang
Abstract: Vision and Language Navigation (VLN) is a challenging task that requires agents to understand instructions and navigate to the destination in a visual environment.One of the key challenges in outdoor VLN is keeping track of which part of the instruction was completed. To alleviate this problem, previous works mainly focus on grounding the natural language to the visual input, but neglecting the crucial role of the agent's spatial position information in the grounding process. In this work, we first explore the substantial effect of spatial position locating on the grounding of outdoor VLN, drawing inspiration from human navigation. In real-world navigation scenarios, before planning a path to the destination, humans typically need to figure out their current location. This observation underscores the pivotal role of spatial localization in the navigation process. In this work, we introduce a novel framework, Locating be for Planning (Loc4Plan), designed to incorporate spatial perception for action planning in outdoor VLN tasks. The main idea behind Loc4Plan is to perform the spatial localization before planning a decision action based on corresponding guidance, which comprises a block-aware spatial locating (BAL) module and a spatial-aware action planning (SAP) module. Specifically, to help the agent perceive its spatial location in the environment, we propose to learn a position predictor that measures how far the agent is from the next intersection for reflecting its position, which is achieved by the BAL module. After the locating process, we propose the SAP module to incorporate spatial information to ground the corresponding guidance and enhance the precision of action planning. Extensive experiments on the Touchdown and map2seq datasets show that the proposed Loc4Plan outperforms the SOTA methods.
Authors: Yamin Sepehri, Pedram Pad, Pascal Frossard, L. Andrea Dunbar
Abstract: The training phase of deep neural networks requires substantial resources and as such is often performed on cloud servers. However, this raises privacy concerns when the training dataset contains sensitive content, e.g., face images. In this work, we propose a method to perform the training phase of a deep learning model on both an edge device and a cloud server that prevents sensitive content being transmitted to the cloud while retaining the desired information. The proposed privacy-preserving method uses adversarial early exits to suppress the sensitive content at the edge and transmits the task-relevant information to the cloud. This approach incorporates noise addition during the training phase to provide a differential privacy guarantee. We extensively test our method on different facial datasets with diverse face attributes using various deep learning architectures, showcasing its outstanding performance. We also demonstrate the effectiveness of privacy preservation through successful defenses against different white-box and deep reconstruction attacks.
Authors: Ziyue Zhang, Mingbao Lin, Shuicheng Yan, Rongrong Ji
Abstract: This paper introduces EasyInv, an easy yet novel approach that significantly advances the field of DDIM Inversion by addressing the inherent inefficiencies and performance limitations of traditional iterative optimization methods. At the core of our EasyInv is a refined strategy for approximating inversion noise, which is pivotal for enhancing the accuracy and reliability of the inversion process. By prioritizing the initial latent state, which encapsulates rich information about the original images, EasyInv steers clear of the iterative refinement of noise items. Instead, we introduce a methodical aggregation of the latent state from the preceding time step with the current state, effectively increasing the influence of the initial latent state and mitigating the impact of noise. We illustrate that EasyInv is capable of delivering results that are either on par with or exceed those of the conventional DDIM Inversion approach, especially under conditions where the model's precision is limited or computational resources are scarce. Concurrently, our EasyInv offers an approximate threefold enhancement regarding inference efficiency over off-the-shelf iterative optimization techniques.
Authors: Yashika Jain, Ali Dabouei, Min Xu
Abstract: Video Anomaly Detection (VAD) automates the identification of unusual events, such as security threats in surveillance videos. In real-world applications, VAD models must effectively operate in cross-domain settings, identifying rare anomalies and scenarios not well-represented in the training data. However, existing cross-domain VAD methods focus on unsupervised learning, resulting in performance that falls short of real-world expectations. Since acquiring weak supervision, i.e., video-level labels, for the source domain is cost-effective, we conjecture that combining it with external unlabeled data has notable potential to enhance cross-domain performance. To this end, we introduce a novel weakly-supervised framework for Cross-Domain Learning (CDL) in VAD that incorporates external data during training by estimating its prediction bias and adaptively minimizing that using the predicted uncertainty. We demonstrate the effectiveness of the proposed CDL framework through comprehensive experiments conducted in various configurations on two large-scale VAD datasets: UCF-Crime and XD-Violence. Our method significantly surpasses the state-of-the-art works in cross-domain evaluations, achieving an average absolute improvement of 19.6% on UCF-Crime and 12.87% on XD-Violence.
Authors: Ruicheng Feng, Chongyi Li, Chen Change Loy
Abstract: Despite the promising progress of face image super-resolution, video face super-resolution remains relatively under-explored. Existing approaches either adapt general video super-resolution networks to face datasets or apply established face image super-resolution models independently on individual video frames. These paradigms encounter challenges either in reconstructing facial details or maintaining temporal consistency. To address these issues, we introduce a novel framework called Kalman-inspired Feature Propagation (KEEP), designed to maintain a stable face prior over time. The Kalman filtering principles offer our method a recurrent ability to use the information from previously restored frames to guide and regulate the restoration process of the current frame. Extensive experiments demonstrate the effectiveness of our method in capturing facial details consistently across video frames. Code and video demo are available at https://jnjaby.github.io/projects/KEEP.
Authors: Yichen Liu, Penghui Du, Yi Liu Quanwei Zhang
Abstract: This paper introduces Multi-Garment Customized Model Generation, a unified framework based on Latent Diffusion Models (LDMs) aimed at addressing the unexplored task of synthesizing images with free combinations of multiple pieces of clothing. The method focuses on generating customized models wearing various targeted outfits according to different text prompts. The primary challenge lies in maintaining the natural appearance of the dressed model while preserving the complex textures of each piece of clothing, ensuring that the information from different garments does not interfere with each other. To tackle these challenges, we first developed a garment encoder, which is a trainable UNet copy with shared weights, capable of extracting detailed features of garments in parallel. Secondly, our framework supports the conditional generation of multiple garments through decoupled multi-garment feature fusion, allowing multiple clothing features to be injected into the backbone network, significantly alleviating conflicts between garment information. Additionally, the proposed garment encoder is a plug-and-play module that can be combined with other extension modules such as IP-Adapter and ControlNet, enhancing the diversity and controllability of the generated models. Extensive experiments demonstrate the superiority of our approach over existing alternatives, opening up new avenues for the task of generating images with multiple-piece clothing combinations
Authors: Chaoyou Fu, Haojia Lin, Zuwei Long, Yunhang Shen, Meng Zhao, Yifan Zhang, Xiong Wang, Di Yin, Long Ma, Xiawu Zheng, Ran He, Rongrong Ji, Yunsheng Wu, Caifeng Shan, Xing Sun
Abstract: The remarkable multimodal capabilities and interactive experience of GPT-4o underscore their necessity in practical applications, yet open-source models rarely excel in both areas. In this paper, we introduce VITA, the first-ever open-source Multimodal Large Language Model (MLLM) adept at simultaneous processing and analysis of Video, Image, Text, and Audio modalities, and meanwhile has an advanced multimodal interactive experience. Starting from Mixtral 8x7B as a language foundation, we expand its Chinese vocabulary followed by bilingual instruction tuning. We further endow the language model with visual and audio capabilities through two-stage multi-task learning of multimodal alignment and instruction tuning. VITA demonstrates robust foundational capabilities of multilingual, vision, and audio understanding, as evidenced by its strong performance across a range of both unimodal and multimodal benchmarks. Beyond foundational capabilities, we have made considerable progress in enhancing the natural multimodal human-computer interaction experience. To the best of our knowledge, we are the first to exploit non-awakening interaction and audio interrupt in MLLM. VITA is the first step for the open-source community to explore the seamless integration of multimodal understanding and interaction. While there is still lots of work to be done on VITA to get close to close-source counterparts, we hope that its role as a pioneer can serve as a cornerstone for subsequent research. Project Page: https://vita-home.github.io.
Authors: Xiaolin Fang, Leslie Pack Kaelbling, Tom\'as Lozano-P\'erez
Abstract: We introduce uncertainty-aware object instance segmentation (UncOS) and demonstrate its usefulness for embodied interactive segmentation. To deal with uncertainty in robot perception, we propose a method for generating a hypothesis distribution of object segmentation. We obtain a set of region-factored segmentation hypotheses together with confidence estimates by making multiple queries of large pre-trained models. This process can produce segmentation results that achieve state-of-the-art performance on unseen object segmentation problems. The output can also serve as input to a belief-driven process for selecting robot actions to perturb the scene to reduce ambiguity. We demonstrate the effectiveness of this method in real-robot experiments. Website: https://sites.google.com/view/embodied-uncertain-seg
Authors: Haider Raza, Mohsin Ali, Vishal Krishna Singh, Agustin Wahjuningrum, Rachel Sarig, Akhilanand Chaurasia
Abstract: Precise identification and detection of the Mental Foramen are crucial in dentistry, impacting procedures such as impacted tooth removal, cyst surgeries, and implants. Accurately identifying this anatomical feature facilitates post-surgery issues and improves patient outcomes. Moreover, this study aims to accelerate dental procedures, elevating patient care and healthcare efficiency in dentistry. This research used Deep Learning methods to accurately detect and segment the Mental Foramen from panoramic radiograph images. Two mask types, circular and square, were used during model training. Multiple segmentation models were employed to identify and segment the Mental Foramen, and their effectiveness was evaluated using diverse metrics. An in-house dataset comprising 1000 panoramic radiographs was created for this study. Our experiments demonstrated that the Classical UNet model performed exceptionally well on the test data, achieving a Dice Coefficient of 0.79 and an Intersection over Union (IoU) of 0.67. Moreover, ResUNet++ and UNet Attention models showed competitive performance, with Dice scores of 0.675 and 0.676, and IoU values of 0.683 and 0.671, respectively. We also investigated transfer learning models with varied backbone architectures, finding LinkNet to produce the best outcomes. In conclusion, our research highlights the efficacy of the classical Unet model in accurately identifying and outlining the Mental Foramen in panoramic radiographs. While vital, this task is comparatively simpler than segmenting complex medical datasets such as brain tumours or skin cancer, given their diverse sizes and shapes. This research also holds value in optimizing dental practice, benefiting practitioners and patients.
Authors: Hao Li, Han Liu, Heinrich von Busch, Robert Grimm, Henkjan Huisman, Angela Tong, David Winkel, Tobias Penzkofer, Ivan Shabunin, Moon Hyung Choi, Qingsong Yang, Dieter Szolar, Steven Shea, Fergus Coakley, Mukesh Harisinghani, Ipek Oguz, Dorin Comaniciu, Ali Kamen, Bin Lou
Abstract: Our hypothesis is that UDA using diffusion-weighted images, generated with a unified model, offers a promising and reliable strategy for enhancing the performance of supervised learning models in multi-site prostate lesion detection, especially when various b-values are present. This retrospective study included data from 5,150 patients (14,191 samples) collected across nine different imaging centers. A novel UDA method using a unified generative model was developed for multi-site PCa detection. This method translates diffusion-weighted imaging (DWI) acquisitions, including apparent diffusion coefficient (ADC) and individual DW images acquired using various b-values, to align with the style of images acquired using b-values recommended by Prostate Imaging Reporting and Data System (PI-RADS) guidelines. The generated ADC and DW images replace the original images for PCa detection. An independent set of 1,692 test cases (2,393 samples) was used for evaluation. The area under the receiver operating characteristic curve (AUC) was used as the primary metric, and statistical analysis was performed via bootstrapping. For all test cases, the AUC values for baseline SL and UDA methods were 0.73 and 0.79 (p<.001), respectively, for PI-RADS>=3, and 0.77 and 0.80 (p<.001) for PI-RADS>=4 PCa lesions. In the 361 test cases under the most unfavorable image acquisition setting, the AUC values for baseline SL and UDA were 0.49 and 0.76 (p<.001) for PI-RADS>=3, and 0.50 and 0.77 (p<.001) for PI-RADS>=4 PCa lesions. The results indicate the proposed UDA with generated images improved the performance of SL methods in multi-site PCa lesion detection across datasets with various b values, especially for images acquired with significant deviations from the PI-RADS recommended DWI protocol (e.g. with an extremely high b-value).
Authors: Dilek M. Yalcinkaya, Khalid Youssef, Bobak Heydari, Janet Wei, Noel Bairey Merz, Robert Judd, Rohan Dharmakumar, Orlando P. Simonetti, Jonathan W. Weinsaft, Subha V. Raman, Behzad Sharif
Abstract: Background. Fully automatic analysis of myocardial perfusion MRI datasets enables rapid and objective reporting of stress/rest studies in patients with suspected ischemic heart disease. Developing deep learning techniques that can analyze multi-center datasets despite limited training data and variations in software and hardware is an ongoing challenge. Methods. Datasets from 3 medical centers acquired at 3T (n = 150 subjects) were included: an internal dataset (inD; n = 95) and two external datasets (exDs; n = 55) used for evaluating the robustness of the trained deep neural network (DNN) models against differences in pulse sequence (exD-1) and scanner vendor (exD-2). A subset of inD (n = 85) was used for training/validation of a pool of DNNs for segmentation, all using the same spatiotemporal U-Net architecture and hyperparameters but with different parameter initializations. We employed a space-time sliding-patch analysis approach that automatically yields a pixel-wise "uncertainty map" as a byproduct of the segmentation process. In our approach, a given test case is segmented by all members of the DNN pool and the resulting uncertainty maps are leveraged to automatically select the "best" one among the pool of solutions. Results. The proposed DAUGS analysis approach performed similarly to the established approach on the internal dataset (p = n.s.) whereas it significantly outperformed on the external datasets (p < 0.005 for exD-1 and exD-2). Moreover, the number of image series with "failed" segmentation was significantly lower for the proposed vs. the established approach (4.3% vs. 17.1%, p < 0.0005). Conclusions. The proposed DAUGS analysis approach has the potential to improve the robustness of deep learning methods for segmentation of multi-center stress perfusion datasets with variations in the choice of pulse sequence, site location or scanner vendor.
Authors: Randall Balestriero, Ahmed Imtiaz Humayun, Richard Baraniuk
Abstract: In this paper, we overview one promising avenue of progress at the mathematical foundation of deep learning: the connection between deep networks and function approximation by affine splines (continuous piecewise linear functions in multiple dimensions). In particular, we will overview work over the past decade on understanding certain geometrical properties of a deep network's affine spline mapping, in particular how it tessellates its input space. As we will see, the affine spline connection and geometrical viewpoint provide a powerful portal through which to view, analyze, and improve the inner workings of a deep network.
Authors: Yiming Chen, Niharika S. D'Souza, Akshith Mandepally, Patrick Henninger, Satyananda Kashyap, Neerav Karani, Neel Dey, Marcos Zachary, Raed Rizq, Paul Chouinard, Polina Golland, Tanveer F. Syeda-Mahmood
Abstract: Precisely estimating lumen boundaries in intravascular ultrasound (IVUS) is needed for sizing interventional stents to treat deep vein thrombosis (DVT). Unfortunately, current segmentation networks like the UNet lack the precision needed for clinical adoption in IVUS workflows. This arises due to the difficulty of automatically learning accurate lumen contour from limited training data while accounting for the radial geometry of IVUS imaging. We propose the Geo-UNet framework to address these issues via a design informed by the geometry of the lumen contour segmentation task. We first convert the input data and segmentation targets from Cartesian to polar coordinates. Starting from a convUNet feature extractor, we propose a two-task setup, one for conventional pixel-wise labeling and the other for single boundary lumen-contour localization. We directly combine the two predictions by passing the predicted lumen contour through a new activation (named CDFeLU) to filter out spurious pixel-wise predictions. Our unified loss function carefully balances area-based, distance-based, and contour-based penalties to provide near clinical-grade generalization in unseen patient data. We also introduce a lightweight, inference-time technique to enhance segmentation smoothness. The efficacy of our framework on a venous IVUS dataset is shown against state-of-the-art models.
Authors: Yuanpu Cao, Lu Lin, Jinghui Chen
Abstract: Deep learning-based industrial anomaly detection models have achieved remarkably high accuracy on commonly used benchmark datasets. However, the robustness of those models may not be satisfactory due to the existence of adversarial examples, which pose significant threats to the practical deployment of deep anomaly detectors. Recently, it has been shown that diffusion models can be used to purify the adversarial noises and thus build a robust classifier against adversarial attacks. Unfortunately, we found that naively applying this strategy in anomaly detection (i.e., placing a purifier before an anomaly detector) will suffer from a high anomaly miss rate since the purifying process can easily remove both the anomaly signal and the adversarial perturbations, causing the later anomaly detector failed to detect anomalies. To tackle this issue, we explore the possibility of performing anomaly detection and adversarial purification simultaneously. We propose a simple yet effective adversarially robust anomaly detection method, \textit{AdvRAD}, that allows the diffusion model to act both as an anomaly detector and adversarial purifier. We also extend our proposed method for certified robustness to $l_2$ norm bounded perturbations. Through extensive experiments, we show that our proposed method exhibits outstanding (certified) adversarial robustness while also maintaining equally strong anomaly detection performance on par with the state-of-the-art methods on industrial anomaly detection benchmark datasets.
Authors: Yue Dai, Soyeon Caren Han, Wei Liu
Abstract: Automatic Chart Question Answering (ChartQA) is challenging due to the complex distribution of chart elements with patterns of the underlying data not explicitly displayed in charts. To address this challenge, we design a joint multimodal scene graph for charts to explicitly represent the relationships between chart elements and their patterns. Our proposed multimodal scene graph includes a visual graph and a textual graph to jointly capture the structural and semantical knowledge from the chart. This graph module can be easily integrated with different vision transformers as inductive bias. Our experiments demonstrate that incorporating the proposed graph module enhances the understanding of charts' elements' structure and semantics, thereby improving performance on publicly available benchmarks, ChartQA and OpenCQA.
Authors: Uri Berger, Gabriel Stanovsky, Omri Abend, Lea Frermann
Abstract: The task of image captioning has recently been gaining popularity, and with it the complex task of evaluating the quality of image captioning models. In this work, we present the first survey and taxonomy of over 70 different image captioning metrics and their usage in hundreds of papers. We find that despite the diversity of proposed metrics, the vast majority of studies rely on only five popular metrics, which we show to be weakly correlated with human judgements. Instead, we propose EnsembEval -- an ensemble of evaluation methods achieving the highest reported correlation with human judgements across 5 image captioning datasets, showing there is a lot of room for improvement by leveraging a diverse set of metrics.
Authors: Gianluca Carloni, Sotirios A Tsaftaris, Sara Colantonio
Abstract: Due to domain shift, deep learning image classifiers perform poorly when applied to a domain different from the training one. For instance, a classifier trained on chest X-ray (CXR) images from one hospital may not generalize to images from another hospital due to variations in scanner settings or patient characteristics. In this paper, we introduce our CROCODILE framework, showing how tools from causality can foster a model's robustness to domain shift via feature disentanglement, contrastive learning losses, and the injection of prior knowledge. This way, the model relies less on spurious correlations, learns the mechanism bringing from images to prediction better, and outperforms baselines on out-of-distribution (OOD) data. We apply our method to multi-label lung disease classification from CXRs, utilizing over 750000 images from four datasets. Our bias-mitigation method improves domain generalization and fairness, broadening the applicability and reliability of deep learning models for a safer medical image analysis. Find our code at: https://github.com/gianlucarloni/crocodile.
Authors: Pietro Morerio, Ruggero Ragonesi, Vittorio Murino
Abstract: Deep Neural Networks are well known for efficiently fitting training data, yet experiencing poor generalization capabilities whenever some kind of bias dominates over the actual task labels, resulting in models learning "shortcuts". In essence, such models are often prone to learn spurious correlations between data and labels. In this work, we tackle the problem of learning from biased data in the very realistic unsupervised scenario, i.e., when the bias is unknown. This is a much harder task as compared to the supervised case, where auxiliary, bias-related annotations, can be exploited in the learning process. This paper proposes a novel 2-stage learning pipeline featuring a data augmentation strategy able to regularize the training. First, biased/unbiased samples are identified by training over-biased models. Second, such subdivision (typically noisy) is exploited within a data augmentation framework, properly combining the original samples while learning mixing parameters, which has a regularization effect. Experiments on synthetic and realistic biased datasets show state-of-the-art classification accuracy, outperforming competing methods, ultimately proving robust performance on both biased and unbiased examples. Notably, being our training method totally agnostic to the level of bias, it also positively affects performance for any, even apparently unbiased, dataset, thus improving the model generalization regardless of the level of bias (or its absence) in the data.
Authors: Kaixin Liu, Huixin Xiong, Bingyu Duan, Zexuan Cheng, Xinyu Zhou, Wanqian Zhang, Xiangyu Zhang
Abstract: In the domain of cloud-based deep learning, the imperative for external computational resources coexists with acute privacy concerns, particularly identity leakage. To address this challenge, we introduce XNN and XNN-d, pioneering methodologies that infuse neural network features with randomized perturbations, striking a harmonious balance between utility and privacy. XNN, designed for the training phase, ingeniously blends random permutation with matrix multiplication techniques to obfuscate feature maps, effectively shielding private data from potential breaches without compromising training integrity. Concurrently, XNN-d, devised for the inference phase, employs adversarial training to integrate generative adversarial noise. This technique effectively counters black-box access attacks aimed at identity extraction, while a distilled face recognition network adeptly processes the perturbed features, ensuring accurate identification. Our evaluation demonstrates XNN's effectiveness, significantly outperforming existing methods in reducing identity leakage while maintaining a high model accuracy.
Authors: Siyue Teng (University of Bristol), Yuxuan Jiang (University of Bristol), Ge Gao (University of Bristol), Fan Zhang (University of Bristol), Thomas Davis (Visionular Inc), Zoe Liu (Visionular Inc), David Bull (University of Bristol)
Abstract: Recent advances in video compression have seen significant coding performance improvements with the development of new standards and learning-based video codecs. However, most of these works focus on application scenarios that allow a certain amount of system delay (e.g., Random Access mode in MPEG codecs), which is not always acceptable for live delivery. This paper conducts a comparative study of state-of-the-art conventional and learned video coding methods based on a low delay configuration. Specifically, this study includes two MPEG standard codecs (H.266/VVC VTM and JVET ECM), two AOM codecs (AV1 libaom and AVM), and two recent neural video coding models (DCVC-DC and DCVC-FM). To allow a fair and meaningful comparison, the evaluation was performed on test sequences defined in the AOM and MPEG common test conditions in the YCbCr 4:2:0 color space. The evaluation results show that the JVET ECM codecs offer the best overall coding performance among all codecs tested, with a 16.1% (based on PSNR) average BD-rate saving over AOM AVM, and 11.0% over DCVC-FM. We also observed inconsistent performance with the learned video codecs, DCVC-DC and DCVC-FM, for test content with large background motions.
Authors: Yoga Sri Varshan V, Hitesh Gupta Kattamuri, Subin Sahayam, Umarani Jayaraman
Abstract: Optic disc and cup segmentation helps in the diagnosis of glaucoma, myocardial infarction, and diabetic retinopathy. Most deep learning methods developed to perform segmentation tasks are built on top of a U-Net-based model architecture. Nevertheless, U-Net and its variants have a tendency to over-segment/ under-segment the required regions of interest. Since the most important outcome is the value of cup-to-disc ratio and not the segmented regions themselves, we are more concerned about the boundaries rather than the regions under the boundaries. This makes learning edges important as compared to learning the regions. In the proposed work, the authors aim to extract both edges of the optic disc and cup from the ground truth using a Laplacian filter. Next, edges are reconstructed to obtain an edge ground truth in addition to the optic disc-cup ground truth. Utilizing both ground truths, the authors study several U-Net and its variant architectures with and without optic disc and cup edges as target, along with the optic disc-cup ground truth for segmentation. The authors have used the REFUGE benchmark dataset and the Drishti-GS dataset to perform the study, and the results are tabulated for the dice and the Hausdorff distance metrics. In the case of the REFUGE dataset, the optic disc mean dice score has improved from 0.7425 to 0.8859 while the mean Hausdorff distance has reduced from 6.5810 to 3.0540 for the baseline U-Net model. Similarly, the optic cup mean dice score has improved from 0.6970 to 0.8639 while the mean Hausdorff distance has reduced from 5.2340 to 2.6323 for the same model. Similar improvement has been observed for the Drishti-GS dataset as well. Compared to the baseline U-Net and its variants (i.e) the Attention U-Net and the U-Net++, the models that learn integrated edges along with the optic disc and cup regions performed well in both validation and testing datasets.
Authors: Ruben Vink, Anna Vilanova, Maxime Chamberland
Abstract: One of the unspoken challenges of tractography is choosing the right parameters for a given dataset or bundle. In order to tackle this challenge, we explore the multi-dimensional parameter space of tractography using streamline-specific parameters (SSP). We 1) validate a state-of-the-art probabilistic tracking method using per-streamline parameters on synthetic data, and 2) show how we can gain insights into the parameter space by focusing on streamline acceptance using real-world data. We demonstrate the potential added value of SSP to the current state of tractography by showing how SSP can be used to reveal patterns in the parameter space.
Authors: Shouyue Liu, Jinkui Hao, Yonghuai Liu, Huazhu Fu, Xinyu Guo, Shuting Zhang, Yitian Zhao
Abstract: Early detection of dementia, such as Alzheimer's disease (AD) or mild cognitive impairment (MCI), is essential to enable timely intervention and potential treatment. Accurate detection of AD/MCI is challenging due to the high complexity, cost, and often invasive nature of current diagnostic techniques, which limit their suitability for large-scale population screening. Given the shared embryological origins and physiological characteristics of the retina and brain, retinal imaging is emerging as a potentially rapid and cost-effective alternative for the identification of individuals with or at high risk of AD. In this paper, we present a novel PolarNet+ that uses retinal optical coherence tomography angiography (OCTA) to discriminate early-onset AD (EOAD) and MCI subjects from controls. Our method first maps OCTA images from Cartesian coordinates to polar coordinates, allowing approximate sub-region calculation to implement the clinician-friendly early treatment of diabetic retinopathy study (ETDRS) grid analysis. We then introduce a multi-view module to serialize and analyze the images along three dimensions for comprehensive, clinically useful information extraction. Finally, we abstract the sequence embedding into a graph, transforming the detection task into a general graph classification problem. A regional relationship module is applied after the multi-view module to excavate the relationship between the sub-regions. Such regional relationship analyses validate known eye-brain links and reveal new discriminative patterns.
Authors: Youqian Zhang, Michael Cheung, Chunxi Yang, Xinwei Zhai, Zitong Shen, Xinyu Ji, Eugene Y. Fu, Sze-Yiu Chau, Xiapu Luo
Abstract: Numerous safety- or security-critical systems depend on cameras to perceive their surroundings, further allowing artificial intelligence (AI) to analyze the captured images to make important decisions. However, a concerning attack vector has emerged, namely, electromagnetic waves, which pose a threat to the integrity of these systems. Such attacks enable attackers to manipulate the images remotely, leading to incorrect AI decisions, e.g., autonomous vehicles missing detecting obstacles ahead resulting in collisions. The lack of understanding regarding how different systems react to such attacks poses a significant security risk. Furthermore, no effective solutions have been demonstrated to mitigate this threat. To address these gaps, we modeled the attacks and developed a simulation method for generating adversarial images. Through rigorous analysis, we confirmed that the effects of the simulated adversarial images are indistinguishable from those from real attacks. This method enables researchers and engineers to rapidly assess the susceptibility of various AI vision applications to these attacks, without the need for constructing complicated attack devices. In our experiments, most of the models demonstrated vulnerabilities to these attacks, emphasizing the need to enhance their robustness. Fortunately, our modeling and simulation method serves as a stepping stone toward developing more resilient models. We present a pilot study on adversarial training to improve their robustness against attacks, and our results demonstrate a significant improvement by recovering up to 91% performance, offering a promising direction for mitigating this threat.
Authors: Marius Bock, Kristof Van Laerhoven, Michael Moeller
Abstract: As wearable-based data annotation remains, to date, a tedious, time-consuming task requiring researchers to dedicate substantial time, benchmark datasets within the field of Human Activity Recognition in lack richness and size compared to datasets available within related fields. Recently, vision foundation models such as CLIP have gained significant attention, helping the vision community advance in finding robust, generalizable feature representations. With the majority of researchers within the wearable community relying on vision modalities to overcome the limited expressiveness of wearable data and accurately label their to-be-released benchmark datasets offline, we propose a novel, clustering-based annotation pipeline to significantly reduce the amount of data that needs to be annotated by a human annotator. We show that using our approach, the annotation of centroid clips suffices to achieve average labelling accuracies close to 90% across three publicly available HAR benchmark datasets. Using the weakly annotated datasets, we further demonstrate that we can match the accuracy scores of fully-supervised deep learning classifiers across all three benchmark datasets. Code as well as supplementary figures and results are publicly downloadable via github.com/mariusbock/weak_har.
Authors: Arne Gevaert, Axel-Jan Rousseau, Thijs Becker, Dirk Valkenborg, Tijl De Bie, Yvan Saeys
Abstract: Feature attribution maps are a popular approach to highlight the most important pixels in an image for a given prediction of a model. Despite a recent growth in popularity and available methods, little attention is given to the objective evaluation of such attribution maps. Building on previous work in this domain, we investigate existing metrics and propose new variants of metrics for the evaluation of attribution maps. We confirm a recent finding that different attribution metrics seem to measure different underlying concepts of attribution maps, and extend this finding to a larger selection of attribution metrics. We also find that metric results on one dataset do not necessarily generalize to other datasets, and methods with desirable theoretical properties such as DeepSHAP do not necessarily outperform computationally cheaper alternatives. Based on these findings, we propose a general benchmarking approach to identify the ideal feature attribution method for a given use case. Implementations of attribution metrics and our experiments are available online.
Authors: Nikolai Kalischek, Torben Peters, Jan D. Wegner, Konrad Schindler
Abstract: Probabilistic denoising diffusion models (DDMs) have set a new standard for 2D image generation. Extending DDMs for 3D content creation is an active field of research. Here, we propose TetraDiffusion, a diffusion model that operates on a tetrahedral partitioning of 3D space to enable efficient, high-resolution 3D shape generation. Our model introduces operators for convolution and transpose convolution that act directly on the tetrahedral partition, and seamlessly includes additional attributes such as color. Remarkably, TetraDiffusion enables rapid sampling of detailed 3D objects in nearly real-time with unprecedented resolution. It's also adaptable for generating 3D shapes conditioned on 2D images. Compared to existing 3D mesh diffusion techniques, our method is up to 200 times faster in inference speed, works on standard consumer hardware, and delivers superior results.
Authors: Lu Sang, Abhishek Saroha, Maolin Gao, Daniel Cremers
Abstract: Neural implicit representations have become a popular choice for modeling surfaces due to their adaptability in resolution and support for complex topology. While previous works have achieved impressive reconstruction quality by training on ground truth point clouds or meshes, they often do not discuss the data acquisition and ignore the effect of input quality and sampling methods during reconstruction. In this paper, we introduce a method that directly digests depth images for the task of high-fidelity 3D reconstruction. To this end, a simple sampling strategy is proposed to generate highly effective training data, by incorporating differentiable geometric features computed directly based on the input depth images with only marginal computational cost. Due to its simplicity, our sampling strategy can be easily incorporated into diverse popular methods, allowing their training process to be more stable and efficient. Despite its simplicity, our method outperforms a range of both classical and learning-based baselines and demonstrates state-of-the-art results in both synthetic and real-world datasets.
Authors: Jinghao Wang, Zhengyu Wen, Xiangtai Li, Zujin Guo, Jingkang Yang, Ziwei Liu
Abstract: Panoptic Scene Graph (PSG) is a challenging task in Scene Graph Generation (SGG) that aims to create a more comprehensive scene graph representation using panoptic segmentation instead of boxes. Compared to SGG, PSG has several challenging problems: pixel-level segment outputs and full relationship exploration (It also considers thing and stuff relation). Thus, current PSG methods have limited performance, which hinders downstream tasks or applications. The goal of this work aims to design a novel and strong baseline for PSG. To achieve that, we first conduct an in-depth analysis to identify the bottleneck of the current PSG models, finding that inter-object pair-wise recall is a crucial factor that was ignored by previous PSG methods. Based on this and the recent query-based frameworks, we present a novel framework: Pair then Relation (Pair-Net), which uses a Pair Proposal Network (PPN) to learn and filter sparse pair-wise relationships between subjects and objects. Moreover, we also observed the sparse nature of object pairs for both Motivated by this, we design a lightweight Matrix Learner within the PPN, which directly learns pair-wised relationships for pair proposal generation. Through extensive ablation and analysis, our approach significantly improves upon leveraging the segmenter solid baseline. Notably, our method achieves over 10\% absolute gains compared to our baseline, PSGFormer. The code of this paper is publicly available at https://github.com/king159/Pair-Net.
Authors: Nakyeong Yang, Minsung Kim, Seunghyun Yoon, Joongbo Shin, Kyomin Jung
Abstract: With the explosion of multimedia content, video moment retrieval (VMR), which aims to detect a video moment that matches a given text query from a video, has been studied intensively as a critical problem. However, the existing VMR framework evaluates video moment retrieval performance, assuming that a video is given, which may not reveal whether the models exhibit overconfidence in the falsely given video. In this paper, we propose the MVMR (Massive Videos Moment Retrieval for Faithfulness Evaluation) task that aims to retrieve video moments within a massive video set, including multiple distractors, to evaluate the faithfulness of VMR models. For this task, we suggest an automated massive video pool construction framework to categorize negative (distractors) and positive (false-negative) video sets using textual and visual semantic distance verification methods. We extend existing VMR datasets using these methods and newly construct three practical MVMR datasets. To solve the task, we further propose a strong informative sample-weighted learning method, CroCs, which employs two contrastive learning mechanisms: (1) weakly-supervised potential negative learning and (2) cross-directional hard-negative learning. Experimental results on the MVMR datasets reveal that existing VMR models are easily distracted by the misinformation (distractors), whereas our model shows significantly robust performance, demonstrating that CroCs is essential to distinguishing positive moments against distractors. Our code and datasets are publicly available: https://github.com/yny0506/Massive-Videos-Moment-Retrieval.
URLs: https://github.com/yny0506/Massive-Videos-Moment-Retrieval.
Authors: Dmitrii Korzh, Mikhail Pautov, Olga Tsymboi, Ivan Oseledets
Abstract: Randomized smoothing is the state-of-the-art approach to construct image classifiers that are provably robust against additive adversarial perturbations of bounded magnitude. However, it is more complicated to construct reasonable certificates against semantic transformation (e.g., image blurring, translation, gamma correction) and their compositions. In this work, we propose \emph{General Lipschitz (GL),} a new framework to certify neural networks against composable resolvable semantic perturbations. Within the framework, we analyze transformation-dependent Lipschitz-continuity of smoothed classifiers w.r.t. transformation parameters and derive corresponding robustness certificates. Our method performs comparably to state-of-the-art approaches on the ImageNet dataset.
Authors: Tingkai Liu, Yunzhe Tao, Haogeng Liu, Qihang Fan, Ding Zhou, Huaibo Huang, Ran He, Hongxia Yang
Abstract: We present a novel human annotated dataset for evaluating the ability for visual-language models to generate both short and long descriptions for real-world video clips, termed DeVAn (Dense Video Annotation). The dataset contains 8.5K YouTube video clips of 20-60 seconds in duration and covers a wide range of topics and interests. Each video clip is independently annotated by 5 human annotators, producing both captions (1 sentence) and summaries (3-10 sentences). Given any video selected from the dataset and its corresponding ASR information, we evaluate visuallanguage models on either caption or summary generation that is grounded in both the visual and auditory content of the video. Additionally, models are also evaluated on caption- and summary-based retrieval tasks, where the summary-based retrieval task requires the identification of a target video given excerpts of a given summary. Given the novel nature of the paragraph-length video summarization task, we compared different existing evaluation metrics and their alignment with human preferences and found that model-based evaluation metrics provide more semantically-oriented and human-aligned evaluation. Finally, we benchmarked a wide range of current video-language models on DeVAn, and we aim for DeVAn to serve as a useful evaluation set in the age of large language models and complex multi-modal tasks. Code is available at https: //github.com/TK-21st/DeVAn.
Authors: Shiye Wang, Kaituo Feng, Changsheng Li, Ye Yuan, Guoren Wang
Abstract: Typical Convolutional Neural Networks (ConvNets) depend heavily on large amounts of image data and resort to an iterative optimization algorithm (e.g., SGD or Adam) to learn network parameters, which makes training very time- and resource-intensive. In this paper, we propose a new training paradigm and formulate the parameter learning of ConvNets into a prediction task: given a ConvNet architecture, we observe there exist correlations between image datasets and their corresponding optimal network parameters, and explore if we can learn a hyper-mapping between them to capture the relations, such that we can directly predict the parameters of the network for an image dataset never seen during the training phase. To do this, we put forward a new hypernetwork based model, called PudNet, which intends to learn a mapping between datasets and their corresponding network parameters, and then predicts parameters for unseen data with only a single forward propagation. Moreover, our model benefits from a series of adaptive hyper recurrent units sharing weights to capture the dependencies of parameters among different network layers. Extensive experiments demonstrate that our proposed method achieves good efficacy for unseen image datasets on two kinds of settings: Intra-dataset prediction and Inter-dataset prediction. Our PudNet can also well scale up to large-scale datasets, e.g., ImageNet-1K. It takes 8967 GPU seconds to train ResNet-18 on the ImageNet-1K using GC from scratch and obtain a top-5 accuracy of 44.65%. However, our PudNet costs only 3.89 GPU seconds to predict the network parameters of ResNet-18 achieving comparable performance (44.92%), more than 2,300 times faster than the traditional training paradigm.
Authors: Zhaoning Wang, Ming Li, Chen Chen
Abstract: With the recent development of generative models, Text-to-3D generations have also seen significant growth, opening a door for creating video-game 3D assets from a more general public. Nonetheless, people without any professional 3D editing experience would find it hard to achieve precise control over the 3D generation, especially if there are multiple objects in the prompt, as using text to control often leads to missing objects and imprecise locations. In this paper, we present LucidDreaming as an effective pipeline capable of spatial and numerical control over 3D generation from only textual prompt commands or 3D bounding boxes. Specifically, our research demonstrates that Large Language Models (LLMs) possess 3D spatial awareness and can effectively translate textual 3D information into precise 3D bounding boxes. We leverage LLMs to get individual object information and their 3D bounding boxes as the initial step of our process. Then with the bounding boxes, We further propose clipped ray sampling and object-centric density blob bias to generate 3D objects aligning with the bounding boxes. We show that our method exhibits remarkable adaptability across a spectrum of mainstream Score Distillation Sampling-based 3D generation frameworks and our pipeline can even used to insert objects into an existing NeRF scene. Moreover, we also provide a dataset of prompts with 3D bounding boxes, benchmarking 3D spatial controllability. With extensive qualitative and quantitative experiments, we demonstrate that LucidDreaming achieves superior results in object placement precision and generation fidelity compared to current approaches, while maintaining flexibility and ease of use for non-expert users.
Authors: Wenxiao Cai, Wankou Yang
Abstract: The topic of stitching images with globally natural structures holds paramount significance, with two main goals: alignment and distortion prevention. The existing approaches exhibit the ability to align well, yet fall short in maintaining object structures. In this paper, we endeavour to safeguard the overall OBJect-level structures within images based on Global Similarity Prior (OBJ-GSP), on the basis of good alignment performance. Our approach leverages semantic segmentation models like the family of Segment Anything Model to extract the contours of any objects in a scene. Triangular meshes are employed in image transformation to protect the overall shapes of objects within images. The balance between alignment and distortion prevention is achieved by allowing the object meshes to strike a balance between similarity and projective transformation. We also demonstrate the importance of segmentation in low-altitude aerial image stitching. Additionally, we propose StitchBench, the most comprehensive image stitching benchmark by far. Extensive experimental results demonstrate that OBJ-GSP outperforms existing methods in both alignment and shape preservation. Code and dataset is publicly available at \url{https://github.com/RussRobin/OBJ-GSP}.
Authors: Jean-Nicolas J\'er\'emie, Emmanuel Dauc\'e, Laurent U Perrinet
Abstract: Foveated vision, a trait shared by many animals, including humans, has not been fully utilized in machine learning applications, despite its significant contributions to biological visual function. This study investigates whether retinotopic mapping, a critical component of foveated vision, can enhance image categorization and localization performance when integrated into deep convolutional neural networks (CNNs). Retinotopic mapping was integrated into the inputs of standard off-the-shelf convolutional neural networks (CNNs), which were then retrained on the ImageNet task. As expected, the logarithmic-polar mapping improved the network's ability to handle arbitrary image zooms and rotations, particularly for isolated objects. Surprisingly, the retinotopically mapped network achieved comparable performance in classification. Furthermore, the network demonstrated improved classification localization when the foveated center of the transform was shifted. This replicates a crucial ability of the human visual system that is absent in typical convolutional neural networks (CNNs). These findings suggest that retinotopic mapping may be fundamental to significant preattentive visual processes.
Authors: Haochen Jiang, Yueming Xu, Yihan Zeng, Hang Xu, Wei Zhang, Jianfeng Feng, Li Zhang
Abstract: 3D reconstruction has been widely used in autonomous navigation fields of mobile robotics. However, the former research can only provide the basic geometry structure without the capability of open-world scene understanding, limiting advanced tasks like human interaction and visual navigation. Moreover, traditional 3D scene understanding approaches rely on expensive labeled 3D datasets to train a model for a single task with supervision. Thus, geometric reconstruction with zero-shot scene understanding i.e. Open vocabulary 3D Understanding and Reconstruction, is crucial for the future development of mobile robots. In this paper, we propose OpenOcc, a novel framework unifying the 3D scene reconstruction and open vocabulary understanding with neural radiance fields. We model the geometric structure of the scene with occupancy representation and distill the pre-trained open vocabulary model into a 3D language field via volume rendering for zero-shot inference. Furthermore, a novel semantic-aware confidence propagation (SCP) method has been proposed to relieve the issue of language field representation degeneracy caused by inconsistent measurements in distilled features. Experimental results show that our approach achieves competitive performance in 3D scene understanding tasks, especially for small and long-tail objects.
Authors: Haz Sameen Shahgir, Khondker Salman Sayeed, Abhik Bhattacharjee, Wasi Uddin Ahmad, Yue Dong, Rifat Shahriyar
Abstract: The advent of Vision Language Models (VLM) has allowed researchers to investigate the visual understanding of a neural network using natural language. Beyond object classification and detection, VLMs are capable of visual comprehension and common-sense reasoning. This naturally led to the question: How do VLMs respond when the image itself is inherently unreasonable? To this end, we present IllusionVQA: a diverse dataset of challenging optical illusions and hard-to-interpret scenes to test the capability of VLMs in two distinct multiple-choice VQA tasks - comprehension and soft localization. GPT4V, the best performing VLM, achieves 62.99% accuracy (4-shot) on the comprehension task and 49.7% on the localization task (4-shot and Chain-of-Thought). Human evaluation reveals that humans achieve 91.03% and 100% accuracy in comprehension and localization. We discover that In-Context Learning (ICL) and Chain-of-Thought reasoning substantially degrade the performance of Gemini-Pro in the localization task. Tangentially, we discover a potential weakness in the ICL capabilities of VLMs: they fail to locate optical illusions even when the correct answer is in the context window as a few-shot example.
Authors: Hannah Schieber, Shiyu Li, Niklas Corell, Philipp Beckerle, Julian Kreimeier, Daniel Roth
Abstract: In medical and industrial domains, providing guidance for assembly processes can be critical to ensure efficiency and safety. Errors in assembly can lead to significant consequences such as extended surgery times and prolonged manufacturing or maintenance times in industry. Assembly scenarios can benefit from in-situ augmented reality visualization, i.e., augmentations in close proximity to the target object, to provide guidance, reduce assembly times, and minimize errors. In order to enable in-situ visualization, 6D pose estimation can be leveraged to identify the correct location for an augmentation. Existing 6D pose estimation techniques primarily focus on individual objects and static captures. However, assembly scenarios have various dynamics, including occlusion during assembly and dynamics in the appearance of assembly objects. Existing work focus either on object detection combined with state detection, or focus purely on the pose estimation. To address the challenges of 6D pose estimation in combination with assembly state detection, our approach ASDF builds upon the strengths of YOLOv8, a real-time capable object detection framework. We extend this framework, refine the object pose, and fuse pose knowledge with network-detected pose information. Utilizing our late fusion in our Pose2State module results in refined 6D pose estimation and assembly state detection. By combining both pose and state information, our Pose2State module predicts the final assembly state with precision. The evaluation of our ASDF dataset shows that our Pose2State module leads to an improved assembly state detection and that the improvement of the assembly state further leads to a more robust 6D pose estimation. Moreover, on the GBOT dataset, we outperform the pure deep learning-based network and even outperform the hybrid and pure tracking-based approaches.
Authors: Hongyun Yu, Zhan Qu, Qihang Yu, Jianchuan Chen, Zhonghua Jiang, Zhiwen Chen, Shengyu Zhang, Jimin Xu, Fei Wu, Chengfei Lv, Gang Yu
Abstract: Recent works on audio-driven talking head synthesis using Neural Radiance Fields (NeRF) have achieved impressive results. However, due to inadequate pose and expression control caused by NeRF implicit representation, these methods still have some limitations, such as unsynchronized or unnatural lip movements, and visual jitter and artifacts. In this paper, we propose GaussianTalker, a novel method for audio-driven talking head synthesis based on 3D Gaussian Splatting. With the explicit representation property of 3D Gaussians, intuitive control of the facial motion is achieved by binding Gaussians to 3D facial models. GaussianTalker consists of two modules, Speaker-specific Motion Translator and Dynamic Gaussian Renderer. Speaker-specific Motion Translator achieves accurate lip movements specific to the target speaker through universalized audio feature extraction and customized lip motion generation. Dynamic Gaussian Renderer introduces Speaker-specific BlendShapes to enhance facial detail representation via a latent pose, delivering stable and realistic rendered videos. Extensive experimental results suggest that GaussianTalker outperforms existing state-of-the-art methods in talking head synthesis, delivering precise lip synchronization and exceptional visual quality. Our method achieves rendering speeds of 130 FPS on NVIDIA RTX4090 GPU, significantly exceeding the threshold for real-time rendering performance, and can potentially be deployed on other hardware platforms.
Authors: Weijie Wang, Jichao Zhang, Chang Liu, Xia Li, Xingqian Xu, Humphrey Shi, Nicu Sebe, Bruno Lepri
Abstract: Recently, diffusion models have made significant strides in synthesizing realistic 2D human images based on provided text prompts. Building upon this, researchers have extended 2D text-to-image diffusion models into the 3D domain for generating human textures (UV Maps). However, some important problems about UV Map Generative models are still not solved, i.e., how to generate personalized texture maps for any given face image, and how to define and evaluate the quality of these generated texture maps. To solve the above problems, we introduce a novel method, UVMap-ID, which is a controllable and personalized UV Map generative model. Unlike traditional large-scale training methods in 2D, we propose to fine-tune a pre-trained text-to-image diffusion model which is integrated with a face fusion module for achieving ID-driven customized generation. To support the finetuning strategy, we introduce a small-scale attribute-balanced training dataset, including high-quality textures with labeled text and Face ID. Additionally, we introduce some metrics to evaluate the multiple aspects of the textures. Finally, both quantitative and qualitative analyses demonstrate the effectiveness of our method in controllable and personalized UV Map generation. Code is publicly available via https://github.com/twowwj/UVMap-ID.
Authors: Huy H. Nguyen, Junichi Yamagishi, Isao Echizen
Abstract: This paper investigates the effectiveness of self-supervised pre-trained vision transformers (ViTs) compared to supervised pre-trained ViTs and conventional neural networks (ConvNets) for detecting facial deepfake images and videos. It examines their potential for improved generalization and explainability, especially with limited training data. Despite the success of transformer architectures in various tasks, the deepfake detection community is hesitant to use large ViTs as feature extractors due to their perceived need for extensive data and suboptimal generalization with small datasets. This contrasts with ConvNets, which are already established as robust feature extractors. Additionally, training ViTs from scratch requires significant resources, limiting their use to large companies. Recent advancements in self-supervised learning (SSL) for ViTs, like masked autoencoders and DINOs, show adaptability across diverse tasks and semantic segmentation capabilities. By leveraging SSL ViTs for deepfake detection with modest data and partial fine-tuning, we find comparable adaptability to deepfake detection and explainability via the attention mechanism. Moreover, partial fine-tuning of ViTs is a resource-efficient option.
Authors: Quang-Huy Che, Le-Chuong Nguyen, Vinh-Tiep Nguyen
Abstract: The quest for robust Person re-identification (Re-ID) systems capable of accurately identifying subjects across diverse scenarios remains a formidable challenge in surveillance and security applications. This study presents a novel methodology that significantly enhances Person Re-Identification (Re-ID) by integrating Uncertainty Feature Fusion (UFFM) with Wise Distance Aggregation (WDA). Tested on benchmark datasets - Market-1501, DukeMTMC-ReID, and MSMT17 - our approach demonstrates substantial improvements in Rank-1 accuracy and mean Average Precision (mAP). Specifically, UFFM capitalizes on the power of feature synthesis from multiple images to overcome the limitations imposed by the variability of subject appearances across different views. WDA further refines the process by intelligently aggregating similarity metrics, thereby enhancing the system's ability to discern subtle but critical differences between subjects. The empirical results affirm the superiority of our method over existing approaches, achieving new performance benchmarks across all evaluated datasets.
Authors: Kai Luo, Hao Wu, Kefu Yi, Kailun Yang, Wei Hao, Rongdong Hu
Abstract: As human-machine interaction continues to evolve, the capacity for environmental perception is becoming increasingly crucial. Integrating the two most common types of sensory data, images, and point clouds, can enhance detection accuracy. Currently, there is no existing model capable of detecting an object's position in both point clouds and images while also determining their corresponding relationship. This information is invaluable for human-machine interactions, offering new possibilities for their enhancement. In light of this, this paper introduces an end-to-end Consistency Object Detection (COD) algorithm framework that requires only a single forward inference to simultaneously obtain an object's position in both point clouds and images and establish their correlation. Furthermore, to assess the accuracy of the object correlation between point clouds and images, this paper proposes a new evaluation metric, Consistency Precision (CP). To verify the effectiveness of the proposed framework, an extensive set of experiments has been conducted on the KITTI and DAIR-V2X datasets. The study also explored how the proposed consistency detection method performs on images when the calibration parameters between images and point clouds are disturbed, compared to existing post-processing methods. The experimental results demonstrate that the proposed method exhibits excellent detection performance and robustness, achieving end-to-end consistency detection. The source code will be made publicly available at https://github.com/xifen523/COD.
Authors: Yizhang Jin, Jian Li, Yexin Liu, Tianjun Gu, Kai Wu, Zhengkai Jiang, Muyang He, Bo Zhao, Xin Tan, Zhenye Gan, Yabiao Wang, Chengjie Wang, Lizhuang Ma
Abstract: In the past year, Multimodal Large Language Models (MLLMs) have demonstrated remarkable performance in tasks such as visual question answering, visual understanding and reasoning. However, the extensive model size and high training and inference costs have hindered the widespread application of MLLMs in academia and industry. Thus, studying efficient and lightweight MLLMs has enormous potential, especially in edge computing scenarios. In this survey, we provide a comprehensive and systematic review of the current state of efficient MLLMs. Specifically, we summarize the timeline of representative efficient MLLMs, research state of efficient structures and strategies, and the applications. Finally, we discuss the limitations of current efficient MLLM research and promising future directions. Please refer to our GitHub repository for more details: https://github.com/lijiannuist/Efficient-Multimodal-LLMs-Survey.
URLs: https://github.com/lijiannuist/Efficient-Multimodal-LLMs-Survey.
Authors: Shiyu Xia, Junyu Xiong, Haoyu Dong, Jianbo Zhao, Yuzhang Tian, Mengyu Zhou, Yeye He, Shi Han, Dongmei Zhang
Abstract: This paper explores capabilities of Vision Language Models on spreadsheet comprehension. We propose three self-supervised challenges with corresponding evaluation metrics to comprehensively evaluate VLMs on Optical Character Recognition (OCR), spatial perception, and visual format recognition. Additionally, we utilize the spreadsheet table detection task to assess the overall performance of VLMs by integrating these challenges. To probe VLMs more finely, we propose three spreadsheet-to-image settings: column width adjustment, style change, and address augmentation. We propose variants of prompts to address the above tasks in different settings. Notably, to leverage the strengths of VLMs in understanding text rather than two-dimensional positioning, we propose to decode cell values on the four boundaries of the table in spreadsheet boundary detection. Our findings reveal that VLMs demonstrate promising OCR capabilities but produce unsatisfactory results due to cell omission and misalignment, and they notably exhibit insufficient spatial and format recognition skills, motivating future work to enhance VLMs' spreadsheet data comprehension capabilities using our methods to generate extensive spreadsheet-image pairs in various settings.
Authors: Mariia Pushkareva, Yuri Feldman, Csaba Domokos, Kilian Rambach, Dotan Di Castro
Abstract: Radar sensors are low cost, long-range, and weather-resilient. Therefore, they are widely used for driver assistance functions, and are expected to be crucial for the success of autonomous driving in the future. In many perception tasks only pre-processed radar point clouds are considered. In contrast, radar spectra are a raw form of radar measurements and contain more information than radar point clouds. However, radar spectra are rather difficult to interpret. In this work, we aim to explore the semantic information contained in spectra in the context of automated driving, thereby moving towards better interpretability of radar spectra. To this end, we create a radar spectra-language model, allowing us to query radar spectra measurements for the presence of scene elements using free text. We overcome the scarcity of radar spectra data by matching the embedding space of an existing vision-language model. Finally, we explore the benefit of the learned representation for scene retrieval using radar spectra only, and obtain improvements in free space segmentation and object detection merely by injecting the spectra embedding into a baseline model.
Authors: Huiwon Jang, Dongyoung Kim, Junsu Kim, Jinwoo Shin, Pieter Abbeel, Younggyo Seo
Abstract: Self-supervised learning of image representations by predicting future frames is a promising direction but still remains a challenge. This is because of the under-determined nature of frame prediction; multiple potential futures can arise from a single current frame. To tackle this challenge, in this paper, we revisit the idea of stochastic video generation that learns to capture uncertainty in frame prediction and explore its effectiveness for representation learning. Specifically, we design a framework that trains a stochastic frame prediction model to learn temporal information between frames. Moreover, to learn dense information within each frame, we introduce an auxiliary masked image modeling objective along with a shared decoder architecture. We find this architecture allows for combining both objectives in a synergistic and compute-efficient manner. We demonstrate the effectiveness of our framework on a variety of tasks from video label propagation and vision-based robot learning domains, such as video segmentation, pose tracking, vision-based robotic locomotion, and manipulation tasks. Code is available on the project webpage: https://sites.google.com/view/2024rsp.
Authors: Yuxuan Zhang, Tianheng Cheng, Rui Hu, Lei Liu, Heng Liu, Longjin Ran, Xiaoxin Chen, Wenyu Liu, Xinggang Wang
Abstract: Segment Anything Model (SAM) has attracted widespread attention for its superior interactive segmentation capabilities with visual prompts while lacking further exploration of text prompts. In this paper, we empirically investigate what text prompt encoders (e.g., CLIP or LLM) are good for adapting SAM for referring expression segmentation and introduce the Early Vision-language Fusion-based SAM (EVF-SAM). EVF-SAM is a simple yet effective referring segmentation method which exploits multimodal prompts (i.e., image and text) and comprises a pre-trained vision-language model to generate referring prompts and a SAM model for segmentation. Surprisingly, we observe that: (1) multimodal prompts and (2) vision-language models with early fusion (e.g., BEIT-3) are beneficial for prompting SAM for accurate referring segmentation. Our experiments show that the proposed EVF-SAM based on BEIT-3 can obtain state-of-the-art performance on RefCOCO/+/g for referring expression segmentation and demonstrate the superiority of prompting SAM with early vision-language fusion. In addition, the proposed EVF-SAM with 1.32B parameters achieves remarkably higher performance while reducing nearly 82% of parameters compared to previous SAM methods based on large multimodal models.
Authors: Wenqi Zhang, Zhenglin Cheng, Yuanyu He, Mengna Wang, Yongliang Shen, Zeqi Tan, Guiyang Hou, Mingqian He, Yanna Ma, Weiming Lu, Yueting Zhuang
Abstract: Although most current large multimodal models (LMMs) can already understand photos of natural scenes and portraits, their understanding of abstract images, e.g., charts, maps, or layouts, and visual reasoning capabilities remains quite rudimentary. They often struggle with simple daily tasks, such as reading time from a clock, understanding a flowchart, or planning a route using a road map. In light of this, we design a multi-modal self-instruct, utilizing large language models and their code capabilities to synthesize massive abstract images and visual reasoning instructions across daily scenarios. Our strategy effortlessly creates a multimodal benchmark with 11,193 instructions for eight visual scenarios: charts, tables, simulated maps, dashboards, flowcharts, relation graphs, floor plans, and visual puzzles. \textbf{This benchmark, constructed with simple lines and geometric elements, exposes the shortcomings of most advanced LMMs} like Claude-3.5-Sonnet and GPT-4o in abstract image understanding, spatial relations reasoning, and visual element induction. Besides, to verify the quality of our synthetic data, we fine-tune an LMM using 62,476 synthetic chart, table and road map instructions. The results demonstrate improved chart understanding and map navigation performance, and also demonstrate potential benefits for other visual reasoning tasks. Our code is available at: \url{https://github.com/zwq2018/Multi-modal-Self-instruct}.
Authors: Josh Myers-Dean, Jarek Reynolds, Brian Price, Yifei Fan, Danna Gurari
Abstract: Hierarchical segmentation entails creating segmentations at varying levels of granularity. We introduce the first hierarchical semantic segmentation dataset with subpart annotations for natural images, which we call SPIN (SubPartImageNet). We also introduce two novel evaluation metrics to evaluate how well algorithms capture spatial and semantic relationships across hierarchical levels. We benchmark modern models across three different tasks and analyze their strengths and weaknesses across objects, parts, and subparts. To facilitate community-wide progress, we publicly release our dataset at https://joshmyersdean.github.io/spin/index.html.
Authors: Fangming Cui, Xun Yang, Chao Wu, Liang Xiao, Xinmei Tian
Abstract: Prompt learning represents a promising method for adapting pre-trained vision-language models (VLMs) to various downstream tasks by learning a set of text embeddings. One challenge inherent to these methods is the poor generalization performance due to the invalidity of the learned text embeddings for unseen tasks. A straightforward approach to bridge this gap is to freeze the text embeddings in prompts, which results in a lack of capacity to adapt VLMs for downstream tasks. To address this dilemma, we propose a paradigm called EnPrompt with a novel External Layer (EnLa). Specifically, we propose a textual external layer and learnable visual embeddings for adapting VLMs to downstream tasks. The learnable external layer is built upon valid embeddings of pre-trained CLIP. This design considers the balance of learning capabilities between the two branches. To align the textual and visual features, we propose a novel two-pronged approach: i) we introduce the optimal transport as the discrepancy metric to align the vision and text modalities, and ii) we introduce a novel strengthening feature to enhance the interaction between these two modalities. Four representative experiments (i.e., base-to-novel generalization, few-shot learning, cross-dataset generalization, domain shifts generalization) across 15 datasets demonstrate that our method outperforms the existing prompt learning method.
Authors: Jingwei Zhu, Olaf Wysocki, Christoph Holst, Thomas H. Kolbe
Abstract: Thermal point clouds integrate thermal radiation and laser point clouds effectively. However, the semantic information for the interpretation of building thermal point clouds can hardly be precisely inferred. Transferring the semantics encapsulated in 3D building models at LoD3 has a potential to fill this gap. In this work, we propose a workflow enriching thermal point clouds with the geo-position and semantics of LoD3 building models, which utilizes features of both modalities: The proposed method can automatically co-register the point clouds from different sources and enrich the thermal point cloud in facade-detailed semantics. The enriched thermal point cloud supports thermal analysis and can facilitate the development of currently scarce deep learning models operating directly on thermal point clouds.
Authors: Mingxin Huang, Yuliang Liu, Dingkang Liang, Lianwen Jin, Xiang Bai
Abstract: Recently, there has been significant interest in enhancing the capability of multimodal large language models (MLLMs) to process high-resolution images. Most existing methods focus on adopting a cropping strategy to improve the ability of multimodal large language models to understand image details. However, this cropping operation inevitably causes the segmentation of objects and connected areas, which impairs the MLLM's ability to recognize small or irregularly shaped objects or text. This issue is particularly evident in lightweight MLLMs. Addressing this issue, we propose Mini-Monkey, a lightweight MLLM that incorporates a plug-and-play method called multi-scale adaptive crop strategy (MSAC). Mini-Monkey adaptively generates multi-scale representations, allowing it to select non-segmented objects from various scales. To mitigate the computational overhead introduced by MSAC, we propose a Scale Compression Mechanism (SCM), which effectively compresses image tokens. Mini-Monkey achieves state-of-the-art performance among 2B-parameter MLLMs. It not only demonstrates leading performance on a variety of general multimodal understanding tasks but also shows consistent improvements in document understanding capabilities. On the OCRBench, Mini-Monkey achieves a score of 802, outperforming 8B-parameter state-of-the-art model InternVL2-8B. Besides, our model and training strategy are very efficient, which can be trained with only eight RTX 3090. The code is available at https://github.com/Yuliang-Liu/Monkey.
Authors: Duc Manh Nguyen Dang, Viet Hang Duong, Jia Ching Wang, Nhan Bui Duc
Abstract: In this paper, we propose a new framework called YOWOv3, which is an improved version of YOWOv2, designed specifically for the task of Human Action Detection and Recognition. This framework is designed to facilitate extensive experimentation with different configurations and supports easy customization of various components within the model, reducing efforts required for understanding and modifying the code. YOWOv3 demonstrates its superior performance compared to YOWOv2 on two widely used datasets for Human Action Detection and Recognition: UCF101-24 and AVAv2.2. Specifically, the predecessor model YOWOv2 achieves an mAP of 85.2% and 20.3% on UCF101-24 and AVAv2.2, respectively, with 109.7M parameters and 53.6 GFLOPs. In contrast, our model - YOWOv3, with only 59.8M parameters and 39.8 GFLOPs, achieves an mAP of 88.33% and 20.31% on UCF101-24 and AVAv2.2, respectively. The results demonstrate that YOWOv3 significantly reduces the number of parameters and GFLOPs while still achieving comparable performance.
Authors: Yiming Zhong, Xiaolin Zhang, Yao Zhao, Yunchao Wei
Abstract: Recently, the text-to-3D task has developed rapidly due to the appearance of the SDS method. However, the SDS method always generates 3D objects with poor quality due to the over-smooth issue. This issue is attributed to two factors: 1) the DDPM single-step inference produces poor guidance gradients; 2) the randomness from the input noises and timesteps averages the details of the 3D contents. In this paper, to address the issue, we propose DreamLCM which incorporates the Latent Consistency Model (LCM). DreamLCM leverages the powerful image generation capabilities inherent in LCM, enabling generating consistent and high-quality guidance, i.e., predicted noises or images. Powered by the improved guidance, the proposed method can provide accurate and detailed gradients to optimize the target 3D models. In addition, we propose two strategies to enhance the generation quality further. Firstly, we propose a guidance calibration strategy, utilizing Euler Solver to calibrate the guidance distribution to accelerate 3D models to converge. Secondly, we propose a dual timestep strategy, increasing the consistency of guidance and optimizing 3D models from geometry to appearance in DreamLCM. Experiments show that DreamLCM achieves state-of-the-art results in both generation quality and training efficiency. The code is available at https://github.com/1YimingZhong/DreamLCM.
Authors: Qirui Jiao, Daoyuan Chen, Yilun Huang, Yaliang Li, Ying Shen
Abstract: High-performance Multimodal Large Language Models (MLLMs) rely heavily on data quality. This study introduces a novel dataset named Img-Diff, designed to enhance fine-grained image recognition in MLLMs by leveraging insights from contrastive learning and image difference captioning. By analyzing object differences between similar images, we challenge models to identify both matching and distinct components. We utilize the Stable-Diffusion-XL model and advanced image editing techniques to create pairs of similar images that highlight object replacements. Our methodology includes a Difference Area Generator for object differences identifying, followed by a Difference Captions Generator for detailed difference descriptions. The result is a relatively small but high-quality dataset of "object replacement" samples. We use the the proposed dataset to finetune state-of-the-art (SOTA) MLLMs such as MGM-7B, yielding comprehensive improvements of performance scores over SOTA models that trained with larger-scale datasets, in numerous image difference and Visual Question Answering tasks. For instance, our trained models notably surpass the SOTA models GPT-4V and Gemini on the MMVP benchmark. Besides, we investigate alternative methods for generating image difference data through "object removal" and conduct a thorough evaluation to confirm the dataset's diversity, quality, and robustness, presenting several insights on the synthesis of such a contrastive dataset. To encourage further research and advance the field of multimodal data synthesis and enhancement of MLLMs' fundamental capabilities for image understanding, we release our codes and dataset at https://github.com/modelscope/data-juicer/tree/ImgDiff.
URLs: https://github.com/modelscope/data-juicer/tree/ImgDiff.
Authors: Weijie J. Su
Abstract: To advance deep learning methodologies in the next decade, a theoretical framework for reasoning about modern neural networks is needed. While efforts are increasing toward demystifying why deep learning is so effective, a comprehensive picture remains lacking, suggesting that a better theory is possible. We argue that a future deep learning theory should inherit three characteristics: a \textit{hierarchically} structured network architecture, parameters \textit{iteratively} optimized using stochastic gradient-based methods, and information from the data that evolves \textit{compressively}. As an instantiation, we integrate these characteristics into a graphical model called \textit{neurashed}. This model effectively explains some common empirical patterns in deep learning. In particular, neurashed enables insights into implicit regularization, information bottleneck, and local elasticity. Finally, we discuss how neurashed can guide the development of deep learning theories.
Authors: Gongning Luo, Kuanquan Wang, Jun Liu, Shuo Li, Xinjie Liang, Xiangyu Li, Shaowei Gan, Wei Wang, Suyu Dong, Wenyi Wang, Pengxin Yu, Enyou Liu, Hongrong Wei, Na Wang, Jia Guo, Huiqi Li, Zhao Zhang, Ziwei Zhao, Na Gao, Nan An, Ashkan Pakzad, Bojidar Rangelov, Jiaqi Dou, Song Tian, Zeyu Liu, Yi Wang, Ampatishan Sivalingam, Kumaradevan Punithakumar, Zhaowen Qiu, Xin Gao
Abstract: Efficient automatic segmentation of multi-level (i.e. main and branch) pulmonary arteries (PA) in CTPA images plays a significant role in clinical applications. However, most existing methods concentrate only on main PA or branch PA segmentation separately and ignore segmentation efficiency. Besides, there is no public large-scale dataset focused on PA segmentation, which makes it highly challenging to compare the different methods. To benchmark multi-level PA segmentation algorithms, we organized the first \textbf{P}ulmonary \textbf{AR}tery \textbf{SE}gmentation (PARSE) challenge. On the one hand, we focus on both the main PA and the branch PA segmentation. On the other hand, for better clinical application, we assign the same score weight to segmentation efficiency (mainly running time and GPU memory consumption during inference) while ensuring PA segmentation accuracy. We present a summary of the top algorithms and offer some suggestions for efficient and accurate multi-level PA automatic segmentation. We provide the PARSE challenge as open-access for the community to benchmark future algorithm developments at \url{https://parse2022.grand-challenge.org/Parse2022/}.
Authors: Hongrong Cheng, Miao Zhang, Javen Qinfeng Shi
Abstract: Modern deep neural networks, particularly recent large language models, come with massive model sizes that require significant computational and storage resources. To enable the deployment of modern models on resource-constrained environments and accelerate inference time, researchers have increasingly explored pruning techniques as a popular research direction in neural network compression. However, there is a dearth of up-to-date comprehensive review papers on pruning. To address this issue, in this survey, we provide a comprehensive review of existing research works on deep neural network pruning in a taxonomy of 1) universal/specific speedup, 2) when to prune, 3) how to prune, and 4) fusion of pruning and other compression techniques. We then provide a thorough comparative analysis of eight pairs of contrast settings for pruning and explore emerging topics, including pruning for large language models, large multimodal models, post-training pruning, and different supervision levels for pruning to shed light on the commonalities and differences of existing methods and lay the foundation for further method development. To facilitate future research, we build a curated collection of datasets, networks, and evaluations on different applications. Finally, we provide valuable recommendations on selecting pruning methods and prospect several promising research directions. We build a repository at https://github.com/hrcheng1066/awesome-pruning.
Authors: Jing Xu
Abstract: The online shopping behavior has the characteristics of rich granularity dimension and data sparsity and previous researches on user behavior prediction did not seriously discuss feature selection and ensemble design. In this paper, we proposed a SE-Stacking model based on information fusion and ensemble learning for user purchase behavior prediction. After successfully utilizing the ensemble feature selection method to screen purchase-related factors, we used the Stacking algorithm for user purchase behavior prediction. In our efforts to avoid the deviation of prediction results, we optimized the model by selecting ten different kinds of models as base learners and modifying relevant parameters specifically for them. The experiments conducted on a publicly-available dataset shows that the SE-Stacking model can achieve a 98.40% F1-score, about 0.09% higher than the optimal base models. The SE-Stacking model not only has a good application in the prediction of user purchase behavior but also has practical value combining with the actual e-commerce scene. At the same time, it has important significance for academic research and the development of this field.
Authors: Jing Xu
Abstract: Chest X-ray is one of the most common radiological examination types for the diagnosis of chest diseases. Nowadays, the automatic classification technology of radiological images has been widely used in clinical diagnosis and treatment plans. However, each disease has its own different response characteristic receptive field region, which is the main challenge for chest disease classification tasks. Besides, the imbalance of sample data categories further increases the difficulty of tasks. To solve these problems, we propose a new multi-label chest disease image classification scheme based on a multi-scale attention network. In this scheme, multi-scale information is iteratively fused to focus on regions with a high probability of disease, to effectively mine more meaningful information from data, and the classification performance can be improved only by image level annotation. We also designed a new loss function to improve the rationality of visual perception and the performance of multi-label image classification by forcing the consistency of attention regions before and after image transformation. A comprehensive experiment was carried out on the public Chest X-Ray14 and CheXpert datasets to achieve state of the art results, which verified the effectiveness of this method in chest X-ray image classification.
Authors: Tao Huang, Guangqi Jiang, Yanjie Ze, Huazhe Xu
Abstract: Learning rewards from expert videos offers an affordable and effective solution to specify the intended behaviors for reinforcement learning (RL) tasks. In this work, we propose Diffusion Reward, a novel framework that learns rewards from expert videos via conditional video diffusion models for solving complex visual RL problems. Our key insight is that lower generative diversity is exhibited when conditioning diffusion on expert trajectories. Diffusion Reward is accordingly formalized by the negative of conditional entropy that encourages productive exploration of expert behaviors. We show the efficacy of our method over robotic manipulation tasks in both simulation platforms and the real world with visual input. Moreover, Diffusion Reward can even solve unseen tasks successfully and effectively, largely surpassing baseline methods. Project page and code: https://diffusion-reward.github.io.
Authors: Mark Karlov, Ali Abedi, Shehroz S. Khan
Abstract: Exercise-based rehabilitation programs have proven to be effective in enhancing the quality of life and reducing mortality and rehospitalization rates. AI-driven virtual rehabilitation, which allows patients to independently complete exercises at home, utilizes AI algorithms to analyze exercise data, providing feedback to patients and updating clinicians on their progress. These programs commonly prescribe a variety of exercise types, leading to a distinct challenge in rehabilitation exercise assessment datasets: while abundant in overall training samples, these datasets often have a limited number of samples for each individual exercise type. This disparity hampers the ability of existing approaches to train generalizable models with such a small sample size per exercise type. Addressing this issue, this paper introduces a novel supervised contrastive learning framework with hard and soft negative samples that effectively utilizes the entire dataset to train a single model applicable to all exercise types. This model, with a Spatial-Temporal Graph Convolutional Network (ST-GCN) architecture, demonstrated enhanced generalizability across exercises and a decrease in overall complexity. Through extensive experiments on three publicly available rehabilitation exercise assessment datasets, UI-PRMD, IRDS, and KIMORE, our method has proven to surpass existing methods, setting a new benchmark in rehabilitation exercise quality assessment.
Authors: Arsham Gholamzadeh Khoee, Yinan Yu, Robert Feldt
Abstract: Deep neural networks (DNNs) have revolutionized artificial intelligence but often lack performance when faced with out-of-distribution (OOD) data, a common scenario due to the inevitable domain shifts in real-world applications. This limitation stems from the common assumption that training and testing data share the same distribution--an assumption frequently violated in practice. Despite their effectiveness with large amounts of data and computational power, DNNs struggle with distributional shifts and limited labeled data, leading to overfitting and poor generalization across various tasks and domains. Meta-learning presents a promising approach by employing algorithms that acquire transferable knowledge across various tasks for fast adaptation, eliminating the need to learn each task from scratch. This survey paper delves into the realm of meta-learning with a focus on its contribution to domain generalization. We first clarify the concept of meta-learning for domain generalization and introduce a novel taxonomy based on the feature extraction strategy and the classifier learning methodology, offering a granular view of methodologies. Additionally, we present a decision graph to assist readers in navigating the taxonomy based on data availability and domain shifts, enabling them to select and develop a proper model tailored to their specific problem requirements. Through an exhaustive review of existing methods and underlying theories, we map out the fundamentals of the field. Our survey provides practical insights and an informed discussion on promising research directions.
Authors: Homanga Bharadhwaj, Roozbeh Mottaghi, Abhinav Gupta, Shubham Tulsiani
Abstract: We seek to learn a generalizable goal-conditioned policy that enables zero-shot robot manipulation: interacting with unseen objects in novel scenes without test-time adaptation. While typical approaches rely on a large amount of demonstration data for such generalization, we propose an approach that leverages web videos to predict plausible interaction plans and learns a task-agnostic transformation to obtain robot actions in the real world. Our framework,Track2Act predicts tracks of how points in an image should move in future time-steps based on a goal, and can be trained with diverse videos on the web including those of humans and robots manipulating everyday objects. We use these 2D track predictions to infer a sequence of rigid transforms of the object to be manipulated, and obtain robot end-effector poses that can be executed in an open-loop manner. We then refine this open-loop plan by predicting residual actions through a closed loop policy trained with a few embodiment-specific demonstrations. We show that this approach of combining scalably learned track prediction with a residual policy requiring minimal in-domain robot-specific data enables diverse generalizable robot manipulation, and present a wide array of real-world robot manipulation results across unseen tasks, objects, and scenes. https://homangab.github.io/track2act/
Authors: Johnny Jingze Li, Vivek Kurien George, Gabriel A. Silva
Abstract: We introduce a novel yet straightforward neural network initialization scheme that modifies conventional methods like Xavier and Kaiming initialization. Inspired by the concept of emergence and leveraging the emergence measures proposed by Li (2023), our method adjusts the layer-wise weight scaling factors to achieve higher emergence values. This enhancement is easy to implement, requiring no additional optimization steps for initialization compared to GradInit. We evaluate our approach across various architectures, including MLP and convolutional architectures for image recognition, and transformers for machine translation. We demonstrate substantial improvements in both model accuracy and training speed, with and without batch normalization. The simplicity, theoretical innovation, and demonstrable empirical advantages of our method make it a potent enhancement to neural network initialization practices. These results suggest a promising direction for leveraging emergence to improve neural network training methodologies. Code is available at: https://github.com/johnnyjingzeli/EmergenceInit.
Authors: Cristian Tommasino, Cristiano Russo, Antonio Maria Rinaldi
Abstract: In pathology, accurate and efficient analysis of Hematoxylin and Eosin (H\&E) slides is crucial for timely and effective cancer diagnosis. Although many deep learning solutions for nuclei instance segmentation and classification exist in the literature, they often entail high computational costs and resource requirements, thus limiting their practical usage in medical applications. To address this issue, we introduce a novel convolutional neural network, NuLite, a U-Net-like architecture designed explicitly on Fast-ViT, a state-of-the-art (SOTA) lightweight CNN. We obtained three versions of our model, NuLite-S, NuLite-M, and NuLite-H, trained on the PanNuke dataset. The experimental results prove that our models equal CellViT (SOTA) in terms of panoptic quality and detection. However, our lightest model, NuLite-S, is 40 times smaller in terms of parameters and about 8 times smaller in terms of GFlops, while our heaviest model is 17 times smaller in terms of parameters and about 7 times smaller in terms of GFlops. Moreover, our model is up to about 8 times faster than CellViT. Lastly, to prove the effectiveness of our solution, we provide a robust comparison of external datasets, namely CoNseP, MoNuSeg, and GlySAC. Our model is publicly available at https://github.com/CosmoIknosLab/NuLite
Authors: Pengcheng Chen, Jin Ye, Guoan Wang, Yanjun Li, Zhongying Deng, Wei Li, Tianbin Li, Haodong Duan, Ziyan Huang, Yanzhou Su, Benyou Wang, Shaoting Zhang, Bin Fu, Jianfei Cai, Bohan Zhuang, Eric J Seibel, Junjun He, Yu Qiao
Abstract: Large Vision-Language Models (LVLMs) are capable of handling diverse data types such as imaging, text, and physiological signals, and can be applied in various fields. In the medical field, LVLMs have a high potential to offer substantial assistance for diagnosis and treatment. Before that, it is crucial to develop benchmarks to evaluate LVLMs' effectiveness in various medical applications. Current benchmarks are often built upon specific academic literature, mainly focusing on a single domain, and lacking varying perceptual granularities. Thus, they face specific challenges, including limited clinical relevance, incomplete evaluations, and insufficient guidance for interactive LVLMs. To address these limitations, we developed the GMAI-MMBench, the most comprehensive general medical AI benchmark with well-categorized data structure and multi-perceptual granularity to date. It is constructed from 285 datasets across 39 medical image modalities, 18 clinical-related tasks, 18 departments, and 4 perceptual granularities in a Visual Question Answering (VQA) format. Additionally, we implemented a lexical tree structure that allows users to customize evaluation tasks, accommodating various assessment needs and substantially supporting medical AI research and applications. We evaluated 50 LVLMs, and the results show that even the advanced GPT-4o only achieves an accuracy of 52%, indicating significant room for improvement. Moreover, we identified five key insufficiencies in current cutting-edge LVLMs that need to be addressed to advance the development of better medical applications. We believe that GMAI-MMBench will stimulate the community to build the next generation of LVLMs toward GMAI.