Authors: Piera Riccio, Georgina Curto, Thomas Hofmann, Nuria Oliver
Abstract: At a time when the influence of generative Artificial Intelligence on visual arts is a highly debated topic, we raise the attention towards a more subtle phenomenon: the algorithmic censorship of artistic nudity online. We analyze the performance of three "Not-Safe-For-Work'' image classifiers on artistic nudity, and empirically uncover the existence of a gender and a stylistic bias, as well as evident technical limitations, especially when only considering visual information. Hence, we propose a multi-modal zero-shot classification approach that improves artistic nudity classification. From our research, we draw several implications that we hope will inform future research on this topic.
Authors: Parth Ganeriwala, Siddhartha Bhattacharyya, Raja Muthalagu
Abstract: Transfer Learning has become one of the standard methods to solve problems to overcome the isolated learning paradigm by utilizing knowledge acquired for one task to solve another related one. However, research needs to be done, to identify the initial steps before inducing transfer learning to applications for further verification and explainablity. In this research, we have performed cross dataset analysis and network architecture repair for the lane detection application in autonomous vehicles. Lane detection is an important aspect of autonomous vehicles driving assistance system. In most circumstances, modern deep-learning-based lane recognition systems are successful, but they struggle with lanes with complex topologies. The proposed architecture, ERFCondLaneNet is an enhancement to the CondlaneNet used for lane identification framework to solve the difficulty of detecting lane lines with complex topologies like dense, curved and fork lines. The newly proposed technique was tested on two common lane detecting benchmarks, CULane and CurveLanes respectively, and two different backbones, ResNet and ERFNet. The researched technique with ERFCondLaneNet, exhibited similar performance in comparison to ResnetCondLaneNet, while using 33% less features, resulting in a reduction of model size by 46%.
Authors: Zaid Ilyas, Afsah Saleem, David Suter, Siobhan Reid, John Schousboe, William Leslie, Joshua Lewis, Syed Zulqarnain Gilani
Abstract: Cardiovascular Diseases (CVDs) are the leading cause of death worldwide, taking 17.9 million lives annually. Abdominal Aortic Calcification (AAC) is an established marker for CVD, which can be observed in lateral view Vertebral Fracture Assessment (VFA) scans, usually done for vertebral fracture detection. Early detection of AAC may help reduce the risk of developing clinical CVDs by encouraging preventive measures. Manual analysis of VFA scans for AAC measurement is time consuming and requires trained human assessors. Recently, efforts have been made to automate the process, however, the proposed models are either low in accuracy, lack granular level score prediction, or are too heavy in terms of inference time and memory footprint. Considering all these shortcomings of existing algorithms, we propose 'AACLiteNet', a lightweight deep learning model that predicts both cumulative and granular level AAC scores with high accuracy, and also has a low memory footprint, and computation cost (Floating Point Operations (FLOPs)). The AACLiteNet achieves a significantly improved one-vs-rest average accuracy of 85.94% as compared to the previous best 81.98%, with 19.88 times less computational cost and 2.26 times less memory footprint, making it implementable on portable computing devices.
Authors: Tommie Kerssies, Daan de Geus, Gijs Dubbelman
Abstract: In this report, we present our solution for Track 1 of the 2024 BRAVO Challenge, where a model is trained on Cityscapes and its robustness is evaluated on several out-of-distribution datasets. Our solution leverages the powerful representations learned by vision foundation models, by attaching a simple segmentation decoder to DINOv2 and fine-tuning the entire model. This approach outperforms more complex existing approaches, and achieves 1st place in the challenge. Our code is publicly available at https://github.com/tue-mps/benchmark-vfm-ss.
Authors: Chaitanya Pallerla, Yihong Feng, Casey M. Owens, Ramesh Bahadur Bist, Siavash Mahmoudi, Pouya Sohrabipour, Amirreza Davar, Dongyi Wang
Abstract: Due to intensive genetic selection for rapid growth rates and high broiler yields in recent years, the global poultry industry has faced a challenging problem in the form of woody breast (WB) conditions. This condition has caused significant economic losses as high as $200 million annually, and the root cause of WB has yet to be identified. Human palpation is the most common method of distinguishing a WB from others. However, this method is time-consuming and subjective. Hyperspectral imaging (HSI) combined with machine learning algorithms can evaluate the WB conditions of fillets in a non-invasive, objective, and high-throughput manner. In this study, 250 raw chicken breast fillet samples (normal, mild, severe) were taken, and spatially heterogeneous hardness distribution was first considered when designing HSI processing models. The study not only classified the WB levels from HSI but also built a regression model to correlate the spectral information with sample hardness data. To achieve a satisfactory classification and regression model, a neural network architecture search (NAS) enabled a wide-deep neural network model named NAS-WD, which was developed. In NAS-WD, NAS was first used to automatically optimize the network architecture and hyperparameters. The classification results show that NAS-WD can classify the three WB levels with an overall accuracy of 95%, outperforming the traditional machine learning model, and the regression correlation between the spectral data and hardness was 0.75, which performs significantly better than traditional regression models.
Authors: Mattia Segu, Luigi Piccinelli, Siyuan Li, Luc Van Gool, Fisher Yu, Bernt Schiele
Abstract: The supervision of state-of-the-art multiple object tracking (MOT) methods requires enormous annotation efforts to provide bounding boxes for all frames of all videos, and instance IDs to associate them through time. To this end, we introduce Walker, the first self-supervised tracker that learns from videos with sparse bounding box annotations, and no tracking labels. First, we design a quasi-dense temporal object appearance graph, and propose a novel multi-positive contrastive objective to optimize random walks on the graph and learn instance similarities. Then, we introduce an algorithm to enforce mutually-exclusive connective properties across instances in the graph, optimizing the learned topology for MOT. At inference time, we propose to associate detected instances to tracklets based on the max-likelihood transition state under motion-constrained bi-directional walks. Walker is the first self-supervised tracker to achieve competitive performance on MOT17, DanceTrack, and BDD100K. Remarkably, our proposal outperforms the previous self-supervised trackers even when drastically reducing the annotation requirements by up to 400x.
Authors: Hui En Pang, Shuai Liu, Zhongang Cai, Lei Yang, Tianwei Zhang, Ziwei Liu
Abstract: We present \textbf{Disco4D}, a novel Gaussian Splatting framework for 4D human generation and animation from a single image. Different from existing methods, Disco4D distinctively disentangles clothings (with Gaussian models) from the human body (with SMPL-X model), significantly enhancing the generation details and flexibility. It has the following technical innovations. \textbf{1)} Disco4D learns to efficiently fit the clothing Gaussians over the SMPL-X Gaussians. \textbf{2)} It adopts diffusion models to enhance the 3D generation process, \textit{e.g.}, modeling occluded parts not visible in the input image. \textbf{3)} It learns an identity encoding for each clothing Gaussian to facilitate the separation and extraction of clothing assets. Furthermore, Disco4D naturally supports 4D human animation with vivid dynamics. Extensive experiments demonstrate the superiority of Disco4D on 4D human generation and animation tasks. Our visualizations can be found in \url{https://disco-4d.github.io/}.
Authors: Zehao Wang, Minye Wu, Yixin Cao, Yubo Ma, Meiqi Chen, Tinne Tuytelaars
Abstract: This study presents a novel evaluation framework for the Vision-Language Navigation (VLN) task. It aims to diagnose current models for various instruction categories at a finer-grained level. The framework is structured around the context-free grammar (CFG) of the task. The CFG serves as the basis for the problem decomposition and the core premise of the instruction categories design. We propose a semi-automatic method for CFG construction with the help of Large-Language Models (LLMs). Then, we induct and generate data spanning five principal instruction categories (i.e. direction change, landmark recognition, region recognition, vertical movement, and numerical comprehension). Our analysis of different models reveals notable performance discrepancies and recurrent issues. The stagnation of numerical comprehension, heavy selective biases over directional concepts, and other interesting findings contribute to the development of future language-guided navigation systems.
Authors: Haodong Li, Hao Lu, Ying-Cong Chen
Abstract: Remote photoplethysmography (rPPG) is gaining prominence for its non-invasive approach to monitoring physiological signals using only cameras. Despite its promise, the adaptability of rPPG models to new, unseen domains is hindered due to the environmental sensitivity of physiological signals. To address this, we pioneer the Test-Time Adaptation (TTA) in rPPG, enabling the adaptation of pre-trained models to the target domain during inference, sidestepping the need for annotations or source data due to privacy considerations. Particularly, utilizing only the user's face video stream as the accessible target domain data, the rPPG model is adjusted by tuning on each single instance it encounters. However, 1) TTA algorithms are designed predominantly for classification tasks, ill-suited in regression tasks such as rPPG due to inadequate supervision. 2) Tuning pre-trained models in a single-instance manner introduces variability and instability, posing challenges to effectively filtering domain-relevant from domain-irrelevant features while simultaneously preserving the learned information. To overcome these challenges, we present Bi-TTA, a novel expert knowledge-based Bidirectional Test-Time Adapter framework. Specifically, leveraging two expert-knowledge priors for providing self-supervision, our Bi-TTA primarily comprises two modules: a prospective adaptation (PA) module using sharpness-aware minimization to eliminate domain-irrelevant noise, enhancing the stability and efficacy during the adaptation process, and a retrospective stabilization (RS) module to dynamically reinforce crucial learned model parameters, averting performance degradation caused by overfitting or catastrophic forgetting. To this end, we established a large-scale benchmark for rPPG tasks under TTA protocol. The experimental results demonstrate the significant superiority of our approach over the state-of-the-art.
Authors: Liangyu Zhong, Joachim Sicking, Fabian H\"uger, Hanno Gottschalk
Abstract: Semantic segmentation networks have achieved significant success under the assumption of independent and identically distributed data. However, these networks often struggle to detect anomalies from unknown semantic classes due to the limited set of visual concepts they are typically trained on. To address this issue, anomaly segmentation often involves fine-tuning on outlier samples, necessitating additional efforts for data collection, labeling, and model retraining. Seeking to avoid this cumbersome work, we take a different approach and propose to incorporate Vision-Language (VL) encoders into existing anomaly detectors to leverage the semantically broad VL pre-training for improved outlier awareness. Additionally, we propose a new scoring function that enables data- and training-free outlier supervision via textual prompts. The resulting VL4AD model, which includes max-logit prompt ensembling and a class-merging strategy, achieves competitive performance on widely used benchmark datasets, thereby demonstrating the potential of vision-language models for pixel-wise anomaly detection.
Authors: Xinhang Liu, Yu-Wing Tai, Chi-Keung Tang
Abstract: Cinematographers adeptly capture the essence of the world, crafting compelling visual narratives through intricate camera movements. Witnessing the strides made by large language models in perceiving and interacting with the 3D world, this study explores their capability to control cameras with human language guidance. We introduce ChatCam, a system that navigates camera movements through conversations with users, mimicking a professional cinematographer's workflow. To achieve this, we propose CineGPT, a GPT-based autoregressive model for text-conditioned camera trajectory generation. We also develop an Anchor Determinator to ensure precise camera trajectory placement. ChatCam understands user requests and employs our proposed tools to generate trajectories, which can be used to render high-quality video footage on radiance field representations. Our experiments, including comparisons to state-of-the-art approaches and user studies, demonstrate our approach's ability to interpret and execute complex instructions for camera operation, showing promising applications in real-world production settings.
Authors: Jay Zoellin, Colin Merk, Mischa Buob, Amr Saad, Samuel Giesser, Tahm Spitznagel, Ferhat Turgut, Rui Santos, Yukun Zhou, Sigfried Wagner, Pearse A. Keane, Yih Chung Tham, Delia Cabrera DeBuc, Matthias D. Becker, Gabor M. Somfai
Abstract: Integrating deep learning into medical imaging is poised to greatly advance diagnostic methods but it faces challenges with generalizability. Foundation models, based on self-supervised learning, address these issues and improve data efficiency. Natural domain foundation models show promise for medical imaging, but systematic research evaluating domain adaptation, especially using self-supervised learning and parameter-efficient fine-tuning, remains underexplored. Additionally, little research addresses the issue of catastrophic forgetting during fine-tuning of foundation models. We adapted the DINOv2 vision transformer for retinal imaging classification tasks using self-supervised learning and generated two novel foundation models termed DINORET and BE DINORET. Publicly available color fundus photographs were employed for model development and subsequent fine-tuning for diabetic retinopathy staging and glaucoma detection. We introduced block expansion as a novel domain adaptation strategy and assessed the models for catastrophic forgetting. Models were benchmarked to RETFound, a state-of-the-art foundation model in ophthalmology. DINORET and BE DINORET demonstrated competitive performance on retinal imaging tasks, with the block expanded model achieving the highest scores on most datasets. Block expansion successfully mitigated catastrophic forgetting. Our few-shot learning studies indicated that DINORET and BE DINORET outperform RETFound in terms of data-efficiency. This study highlights the potential of adapting natural domain vision models to retinal imaging using self-supervised learning and block expansion. BE DINORET offers robust performance without sacrificing previously acquired capabilities. Our findings suggest that these methods could enable healthcare institutions to develop tailored vision models for their patient populations, enhancing global healthcare inclusivity.
Authors: Md Abdullah-Al Kaiser, Sreetama Sarkar, Peter A. Beerel, Akhilesh R. Jaiswal, Gourav Datta
Abstract: Current video-based computer vision (CV) applications typically suffer from high energy consumption due to reading and processing all pixels in a frame, regardless of their significance. While previous works have attempted to reduce this energy by skipping input patches or pixels and using feedback from the end task to guide the skipping algorithm, the skipping is not performed during the sensor read phase. As a result, these methods can not optimize the front-end sensor energy. Moreover, they may not be suitable for real-time applications due to the long latency of modern CV networks that are deployed in the back-end. To address this challenge, this paper presents a custom-designed reconfigurable CMOS image sensor (CIS) system that improves energy efficiency by selectively skipping uneventful regions or rows within a frame during the sensor's readout phase, and the subsequent analog-to-digital conversion (ADC) phase. A novel masking algorithm intelligently directs the skipping process in real-time, optimizing both the front-end sensor and back-end neural networks for applications including autonomous driving and augmented/virtual reality (AR/VR). Our system can also operate in standard mode without skipping, depending on application needs. We evaluate our hardware-algorithm co-design framework on object detection based on BDD100K and ImageNetVID, and gaze estimation based on OpenEDS, achieving up to 53% reduction in front-end sensor energy while maintaining state-of-the-art (SOTA) accuracy.
Authors: Daniel Yang, John J. Leonard, Yogesh Girdhar
Abstract: We introduce SeaSplat, a method to enable real-time rendering of underwater scenes leveraging recent advances in 3D radiance fields. Underwater scenes are challenging visual environments, as rendering through a medium such as water introduces both range and color dependent effects on image capture. We constrain 3D Gaussian Splatting (3DGS), a recent advance in radiance fields enabling rapid training and real-time rendering of full 3D scenes, with a physically grounded underwater image formation model. Applying SeaSplat to the real-world scenes from SeaThru-NeRF dataset, a scene collected by an underwater vehicle in the US Virgin Islands, and simulation-degraded real-world scenes, not only do we see increased quantitative performance on rendering novel viewpoints from the scene with the medium present, but are also able to recover the underlying true color of the scene and restore renders to be without the presence of the intervening medium. We show that the underwater image formation helps learn scene structure, with better depth maps, as well as show that our improvements maintain the significant computational improvements afforded by leveraging a 3D Gaussian representation.
Authors: Konstantinos Papoutsakis, Nikolaos Bakalos, Konstantinos Fragkoulis, Athena Zacharia, Georgia Kapetadimitri, Maria Pateraki
Abstract: This paper introduces a vision-based framework for capturing and understanding human behavior in industrial assembly lines, focusing on car door manufacturing. The framework leverages advanced computer vision techniques to estimate workers' locations and 3D poses and analyze work postures, actions, and task progress. A key contribution is the introduction of the CarDA dataset, which contains domain-relevant assembly actions captured in a realistic setting to support the analysis of the framework for human pose and action analysis. The dataset comprises time-synchronized multi-camera RGB-D videos, motion capture data recorded in a real car manufacturing environment, and annotations for EAWS-based ergonomic risk scores and assembly activities. Experimental results demonstrate the effectiveness of the proposed approach in classifying worker postures and robust performance in monitoring assembly task progress.
Authors: Kartik Jindgar, Grace W. Lindsay
Abstract: In recent years, analysis of remote sensing data has benefited immensely from borrowing techniques from the broader field of computer vision, such as the use of shared models pre-trained on large and diverse datasets. However, satellite imagery has unique features that are not accounted for in traditional computer vision, such as the existence of multiple revisits of the same location. Here, we explore the best way to use revisits in the framework of fine-tuning pre-trained remote sensing models. We focus on an applied research question of relevance to climate change mitigation -- power substation segmentation -- that is representative of applied uses of pre-trained models more generally. Through extensive tests of different multi-temporal input schemes across diverse model architectures, we find that fusing representations from multiple revisits in the model latent space is superior to other methods of using revisits, including as a form of data augmentation. We also find that a SWIN Transformer-based architecture performs better than U-nets and ViT-based models. We verify the generality of our results on a separate building density estimation task.
Authors: David Bertoin, Eduardo Hugo Sanchez, Mehdi Zouitine, Emmanuel Rachelson
Abstract: Despite transformers being considered as the new standard in computer vision, convolutional neural networks (CNNs) still outperform them in low-data regimes. Nonetheless, CNNs often make decisions based on narrow, specific regions of input images, especially when training data is limited. This behavior can severely compromise the model's generalization capabilities, making it disproportionately dependent on certain features that might not represent the broader context of images. While the conditions leading to this phenomenon remain elusive, the primary intent of this article is to shed light on this observed behavior of neural networks. Our research endeavors to prioritize comprehensive insight and to outline an initial response to this phenomenon. In line with this, we introduce Saliency Guided Dropout (SGDrop), a pioneering regularization approach tailored to address this specific issue. SGDrop utilizes attribution methods on the feature map to identify and then reduce the influence of the most salient features during training. This process encourages the network to diversify its attention and not focus solely on specific standout areas. Our experiments across several visual classification benchmarks validate SGDrop's role in enhancing generalization. Significantly, models incorporating SGDrop display more expansive attributions and neural activity, offering a more comprehensive view of input images in contrast to their traditionally trained counterparts.
Authors: Uddhav Bhattarai, Santosh Bhusal, Qin Zhang, Manoj Karkee
Abstract: One of the major challenges for the agricultural industry today is the uncertainty in manual labor availability and the associated cost. Automated flower and fruit density estimation, localization, and counting could help streamline harvesting, yield estimation, and crop-load management strategies such as flower and fruitlet thinning. This article proposes a deep regression-based network, AgRegNet, to estimate density, count, and location of flower and fruit in tree fruit canopies without explicit object detection or polygon annotation. Inspired by popular U-Net architecture, AgRegNet is a U-shaped network with an encoder-to-decoder skip connection and modified ConvNeXt-T as an encoder feature extractor. AgRegNet can be trained based on information from point annotation and leverages segmentation information and attention modules (spatial and channel) to highlight relevant flower and fruit features while suppressing non-relevant background features. Experimental evaluation in apple flower and fruit canopy images under an unstructured orchard environment showed that AgRegNet achieved promising accuracy as measured by Structural Similarity Index (SSIM), percentage Mean Absolute Error (pMAE) and mean Average Precision (mAP) to estimate flower and fruit density, count, and centroid location, respectively. Specifically, the SSIM, pMAE, and mAP values for flower images were 0.938, 13.7%, and 0.81, respectively. For fruit images, the corresponding values were 0.910, 5.6%, and 0.93. Since the proposed approach relies on information from point annotation, it is suitable for sparsely and densely located objects. This simplified technique will be highly applicable for growers to accurately estimate yields and decide on optimal chemical and mechanical flower thinning practices.
Authors: Md Tanvir Islam, Nasir Rahim, Saeed Anwar, Muhammad Saqib, Sambit Bakshi, Khan Muhammad
Abstract: Reducing the atmospheric haze and enhancing image clarity is crucial for computer vision applications. The lack of real-life hazy ground truth images necessitates synthetic datasets, which often lack diverse haze types, impeding effective haze type classification and dehazing algorithm selection. This research introduces the HazeSpace2M dataset, a collection of over 2 million images designed to enhance dehazing through haze type classification. HazeSpace2M includes diverse scenes with 10 haze intensity levels, featuring Fog, Cloud, and Environmental Haze (EH). Using the dataset, we introduce a technique of haze type classification followed by specialized dehazers to clear hazy images. Unlike conventional methods, our approach classifies haze types before applying type-specific dehazing, improving clarity in real-life hazy images. Benchmarking with state-of-the-art (SOTA) models, ResNet50 and AlexNet achieve 92.75\% and 92.50\% accuracy, respectively, against existing synthetic datasets. However, these models achieve only 80% and 70% accuracy, respectively, against our Real Hazy Testset (RHT), highlighting the challenging nature of our HazeSpace2M dataset. Additional experiments show that haze type classification followed by specialized dehazing improves results by 2.41% in PSNR, 17.14% in SSIM, and 10.2\% in MSE over general dehazers. Moreover, when testing with SOTA dehazing models, we found that applying our proposed framework significantly improves their performance. These results underscore the significance of HazeSpace2M and our proposed framework in addressing atmospheric haze in multimedia processing. Complete code and dataset is available on \href{https://github.com/tanvirnwu/HazeSpace2M} {\textcolor{blue}{\textbf{GitHub}}}.
Authors: Chirag Vashist, Shichong Peng, Ke Li
Abstract: An emerging area of research aims to learn deep generative models with limited training data. Prior generative models like GANs and diffusion models require a lot of data to perform well, and their performance degrades when they are trained on only a small amount of data. A recent technique called Implicit Maximum Likelihood Estimation (IMLE) has been adapted to the few-shot setting, achieving state-of-the-art performance. However, current IMLE-based approaches encounter challenges due to inadequate correspondence between the latent codes selected for training and those drawn during inference. This results in suboptimal test-time performance. We theoretically show a way to address this issue and propose RS-IMLE, a novel approach that changes the prior distribution used for training. This leads to substantially higher quality image generation compared to existing GAN and IMLE-based methods, as validated by comprehensive experiments conducted on nine few-shot image datasets.
Authors: Hanbo Bi, Yingchao Feng, Yongqiang Mao, Jianning Pei, Wenhui Diao, Hongqi Wang, Xian Sun
Abstract: Few-shot Segmentation (FSS) aims to segment the interested objects in the query image with just a handful of labeled samples (i.e., support images). Previous schemes would leverage the similarity between support-query pixel pairs to construct the pixel-level semantic correlation. However, in remote sensing scenarios with extreme intra-class variations and cluttered backgrounds, such pixel-level correlations may produce tremendous mismatches, resulting in semantic ambiguity between the query foreground (FG) and background (BG) pixels. To tackle this problem, we propose a novel Agent Mining Transformer (AgMTR), which adaptively mines a set of local-aware agents to construct agent-level semantic correlation. Compared with pixel-level semantics, the given agents are equipped with local-contextual information and possess a broader receptive field. At this point, different query pixels can selectively aggregate the fine-grained local semantics of different agents, thereby enhancing the semantic clarity between query FG and BG pixels. Concretely, the Agent Learning Encoder (ALE) is first proposed to erect the optimal transport plan that arranges different agents to aggregate support semantics under different local regions. Then, for further optimizing the agents, the Agent Aggregation Decoder (AAD) and the Semantic Alignment Decoder (SAD) are constructed to break through the limited support set for mining valuable class-specific semantics from unlabeled data sources and the query image itself, respectively. Extensive experiments on the remote sensing benchmark iSAID indicate that the proposed method achieves state-of-the-art performance. Surprisingly, our method remains quite competitive when extended to more common natural scenarios, i.e., PASCAL-5i and COCO-20i.
Authors: Sifan Wu, Amir Khasahmadi, Mor Katz, Pradeep Kumar Jayaraman, Yewen Pu, Karl Willis, Bang Liu
Abstract: Parametric Computer-Aided Design (CAD) is central to contemporary mechanical design. However, it encounters challenges in achieving precise parametric sketch modeling and lacks practical evaluation metrics suitable for mechanical design. We harness the capabilities of pre-trained foundation models, renowned for their successes in natural language processing and computer vision, to develop generative models specifically for CAD. These models are adept at understanding complex geometries and design reasoning, a crucial advancement in CAD technology. In this paper, we propose CadVLM, an end-to-end vision language model for CAD generation. Our approach involves adapting pre-trained foundation models to manipulate engineering sketches effectively, integrating both sketch primitive sequences and sketch images. Extensive experiments demonstrate superior performance on multiple CAD sketch generation tasks such as CAD autocompletion, CAD autoconstraint, and image conditional generation. To our knowledge, this is the first instance of a multimodal Large Language Model (LLM) being successfully applied to parametric CAD generation, representing a pioneering step in the field of computer-aided mechanical design.
Authors: Sandika Biswas, Qianyi Wu, Biplab Banerjee, Hamid Rezatofighi
Abstract: Despite advancements in Neural Implicit models for 3D surface reconstruction, handling dynamic environments with arbitrary rigid, non-rigid, or deformable entities remains challenging. Many template-based methods are entity-specific, focusing on humans, while generic reconstruction methods adaptable to such dynamic scenes often require additional inputs like depth or optical flow or rely on pre-trained image features for reasonable outcomes. These methods typically use latent codes to capture frame-by-frame deformations. In contrast, some template-free methods bypass these requirements and adopt traditional LBS (Linear Blend Skinning) weights for a detailed representation of deformable object motions, although they involve complex optimizations leading to lengthy training times. To this end, as a remedy, this paper introduces TFS-NeRF, a template-free 3D semantic NeRF for dynamic scenes captured from sparse or single-view RGB videos, featuring interactions among various entities and more time-efficient than other LBS-based approaches. Our framework uses an Invertible Neural Network (INN) for LBS prediction, simplifying the training process. By disentangling the motions of multiple entities and optimizing per-entity skinning weights, our method efficiently generates accurate, semantically separable geometries. Extensive experiments demonstrate that our approach produces high-quality reconstructions of both deformable and non-deformable objects in complex interactions, with improved training efficiency compared to existing methods.
Authors: Yi Gu, Yi Lin, Kwang-Ting Cheng, Hao Chen
Abstract: Medical anomaly detection (AD) is crucial in pathological identification and localization. Current methods typically rely on uncertainty estimation in deep ensembles to detect anomalies, assuming that ensemble learners should agree on normal samples while exhibiting disagreement on unseen anomalies in the output space. However, these methods may suffer from inadequate disagreement on anomalies or diminished agreement on normal samples. To tackle these issues, we propose D2UE, a Diversified Dual-space Uncertainty Estimation framework for medical anomaly detection. To effectively balance agreement and disagreement for anomaly detection, we propose Redundancy-Aware Repulsion (RAR), which uses a similarity kernel that remains invariant to both isotropic scaling and orthogonal transformations, explicitly promoting diversity in learners' feature space. Moreover, to accentuate anomalous regions, we develop Dual-Space Uncertainty (DSU), which utilizes the ensemble's uncertainty in input and output spaces. In input space, we first calculate gradients of reconstruction error with respect to input images. The gradients are then integrated with reconstruction outputs to estimate uncertainty for inputs, enabling effective anomaly discrimination even when output space disagreement is minimal. We conduct a comprehensive evaluation of five medical benchmarks with different backbones. Experimental results demonstrate the superiority of our method to state-of-the-art methods and the effectiveness of each component in our framework. Our code is available at https://github.com/Rubiscol/D2UE.
Authors: Yuchen Liang, Yuchuan Tian, Lei Yu, Huao Tang, Jie Hu, Xiangzhong Fang, Hanting Chen
Abstract: The curvature of ODE trajectories in diffusion models hinders their ability to generate high-quality images in a few number of function evaluations (NFE). In this paper, we propose a novel and effective approach to reduce trajectory curvature by utilizing adaptive conditions. By employing a extremely light-weight quantized encoder, our method incurs only an additional 1% of training parameters, eliminates the need for extra regularization terms, yet achieves significantly better sample quality. Our approach accelerates ODE sampling while preserving the downstream task image editing capabilities of SDE techniques. Extensive experiments verify that our method can generate high quality results under extremely limited sampling costs. With only 6 NFE, we achieve 5.14 FID on CIFAR-10, 6.91 FID on FFHQ 64x64 and 3.10 FID on AFHQv2.
Authors: Xun Zhu, Ying Hu, Fanbin Mo, Miao Li, Ji Wu
Abstract: Multi-modal large language models (MLLMs) have shown impressive capabilities as a general-purpose interface for various visual and linguistic tasks. However, building a unified MLLM for multi-task learning in the medical field remains a thorny challenge. To mitigate the tug-of-war problem of multi-modal multi-task optimization, recent advances primarily focus on improving the LLM components, while neglecting the connector that bridges the gap between modalities. In this paper, we introduce Uni-Med, a novel medical generalist foundation model which consists of a universal visual feature extraction module, a connector mixture-of-experts (CMoE) module, and an LLM. Benefiting from the proposed CMoE that leverages a well-designed router with a mixture of projection experts at the connector, Uni-Med achieves efficient solution to the tug-of-war problem and can perform six different medical tasks including question answering, visual question answering, report generation, referring expression comprehension, referring expression generation and image classification. To the best of our knowledge, Uni-Med is the first effort to tackle multi-task interference at the connector. Extensive ablation experiments validate the effectiveness of introducing CMoE under any configuration, with up to an average 8% performance gains. We further provide interpretation analysis of the tug-of-war problem from the perspective of gradient optimization and parameter statistics. Compared to previous state-of-the-art medical MLLMs, Uni-Med achieves competitive or superior evaluation metrics on diverse tasks. Code, data and model will be soon available at GitHub.
Authors: Zerun Wang, Liuyu Xiang, Lang Huang, Jiafeng Mao, Ling Xiao, Toshihiko Yamasaki
Abstract: Open-set semi-supervised learning (OSSL) leverages practical open-set unlabeled data, comprising both in-distribution (ID) samples from seen classes and out-of-distribution (OOD) samples from unseen classes, for semi-supervised learning (SSL). Prior OSSL methods initially learned the decision boundary between ID and OOD with labeled ID data, subsequently employing self-training to refine this boundary. These methods, however, suffer from the tendency to overtrust the labeled ID data: the scarcity of labeled data caused the distribution bias between the labeled samples and the entire ID data, which misleads the decision boundary to overfit. The subsequent self-training process, based on the overfitted result, fails to rectify this problem. In this paper, we address the overtrusting issue by treating OOD samples as an additional class, forming a new SSL process. Specifically, we propose SCOMatch, a novel OSSL method that 1) selects reliable OOD samples as new labeled data with an OOD memory queue and a corresponding update strategy and 2) integrates the new SSL process into the original task through our Simultaneous Close-set and Open-set self-training. SCOMatch refines the decision boundary of ID and OOD classes across the entire dataset, thereby leading to improved results. Extensive experimental results show that SCOMatch significantly outperforms the state-of-the-art methods on various benchmarks. The effectiveness is further verified through ablation studies and visualization.
Authors: Jing Bi, Yunlong Tang, Luchuan Song, Ali Vosoughi, Nguyen Nguyen, Chenliang Xu
Abstract: The rapid evolution of egocentric video analysis brings new insights into understanding human activities and intentions from a first-person perspective. Despite this progress, the fragmentation in tasks like action recognition, procedure learning, and moment retrieval, \etc, coupled with inconsistent annotations and isolated model development, hinders a holistic interpretation of video content. In response, we introduce the EAGLE (Egocentric AGgregated Language-video Engine) model and the EAGLE-400K dataset to provide a unified framework that integrates various egocentric video understanding tasks. EAGLE-400K, the \textit{first} large-scale instruction-tuning dataset tailored for egocentric video, features 400K diverse samples to enhance a broad spectrum of tasks from activity recognition to procedure knowledge learning. Moreover, EAGLE, a strong video multimodal large language model (MLLM), is designed to effectively capture both spatial and temporal information. In addition, we propose a set of evaluation metrics designed to facilitate a thorough assessment of MLLM for egocentric video understanding. Our extensive experiments demonstrate EAGLE's superior performance over existing models, highlighting its ability to balance task-specific understanding with holistic video interpretation. With EAGLE, we aim to pave the way for research opportunities and practical applications in real-world scenarios.
Authors: Chao Li, Chen Jiang, Xiaolong Liu, Jun Zhao, Guoxin Wang
Abstract: Generating images with accurately represented text, especially in non-Latin languages, poses a significant challenge for diffusion models. Existing approaches, such as the integration of hint condition diagrams via auxiliary networks (e.g., ControlNet), have made strides towards addressing this issue. However, diffusion models often fall short in tasks requiring controlled text generation, such as specifying particular fonts or producing text in small fonts. In this paper, we introduce a novel approach for multilingual visual text creation, named JoyType, designed to maintain the font style of text during the image generation process. Our methodology begins with assembling a training dataset, JoyType-1M, comprising 1 million pairs of data. Each pair includes an image, its description, and glyph instructions corresponding to the font style within the image. We then developed a text control network, Font ControlNet, tasked with extracting font style information to steer the image generation. To further enhance our model's ability to maintain font style, notably in generating small-font text, we incorporated a multi-layer OCR-aware loss into the diffusion process. This enhancement allows JoyType to direct text rendering using low-level descriptors. Our evaluations, based on both visual and accuracy metrics, demonstrate that JoyType significantly outperforms existing state-of-the-art methods. Additionally, JoyType can function as a plugin, facilitating the creation of varied image styles in conjunction with other stable diffusion models on HuggingFace and CivitAI. Our project is open-sourced on https://jdh-algo.github.io/JoyType/.
Authors: Yida Lin, Bing Xue, Mengjie Zhang, Sam Schofield, Richard Green
Abstract: Manual pruning of radiata pine trees presents significant safety risks due to their substantial height and the challenging terrains in which they thrive. To address these risks, this research proposes the development of a drone-based pruning system equipped with specialized pruning tools and a stereo vision camera, enabling precise detection and trimming of branches. Deep learning algorithms, including YOLO and Mask R-CNN, are employed to ensure accurate branch detection, while the Semi-Global Matching algorithm is integrated to provide reliable distance estimation. The synergy between these techniques facilitates the precise identification of branch locations and enables efficient, targeted pruning. Experimental results demonstrate that the combined implementation of YOLO and SGBM enables the drone to accurately detect branches and measure their distances from the drone. This research not only improves the safety and efficiency of pruning operations but also makes a significant contribution to the advancement of drone technology in the automation of agricultural and forestry practices, laying a foundational framework for further innovations in environmental management.
Authors: Ming Dai, Lingfeng Yang, Yihao Xu, Zhenhua Feng, Wankou Yang
Abstract: Visual grounding is a common vision task that involves grounding descriptive sentences to the corresponding regions of an image. Most existing methods use independent image-text encoding and apply complex hand-crafted modules or encoder-decoder architectures for modal interaction and query reasoning. However, their performance significantly drops when dealing with complex textual expressions. This is because the former paradigm only utilizes limited downstream data to fit the multi-modal feature fusion. Therefore, it is only effective when the textual expressions are relatively simple. In contrast, given the wide diversity of textual expressions and the uniqueness of downstream training data, the existing fusion module, which extracts multimodal content from a visual-linguistic context, has not been fully investigated. In this paper, we present a simple yet robust transformer-based framework, SimVG, for visual grounding. Specifically, we decouple visual-linguistic feature fusion from downstream tasks by leveraging existing multimodal pre-trained models and incorporating additional object tokens to facilitate deep integration of downstream and pre-training tasks. Furthermore, we design a dynamic weight-balance distillation method in the multi-branch synchronous learning process to enhance the representation capability of the simpler branch. This branch only consists of a lightweight MLP, which simplifies the structure and improves reasoning speed. Experiments on six widely used VG datasets, i.e., RefCOCO/+/g, ReferIt, Flickr30K, and GRefCOCO, demonstrate the superiority of SimVG. Finally, the proposed method not only achieves improvements in efficiency and convergence speed but also attains new state-of-the-art performance on these benchmarks. Codes and models will be available at \url{https://github.com/Dmmm1997/SimVG}.
Authors: Felix Limanta, Kuniaki Uto, Koichi Shinoda
Abstract: This paper proposes CAMOT, a simple camera angle estimator for multi-object tracking to tackle two problems: 1) occlusion and 2) inaccurate distance estimation in the depth direction. Under the assumption that multiple objects are located on a flat plane in each video frame, CAMOT estimates the camera angle using object detection. In addition, it gives the depth of each object, enabling pseudo-3D MOT. We evaluated its performance by adding it to various 2D MOT methods on the MOT17 and MOT20 datasets and confirmed its effectiveness. Applying CAMOT to ByteTrack, we obtained 63.8% HOTA, 80.6% MOTA, and 78.5% IDF1 in MOT17, which are state-of-the-art results. Its computational cost is significantly lower than the existing deep-learning-based depth estimators for tracking.
Authors: Jiaming Liu, Linghe Kong, Yue Wu, Maoguo Gong, Hao Li, Qiguang Miao, Wenping Ma, Can Qin
Abstract: Existing 3D mask learning methods encounter performance bottlenecks under limited data, and our objective is to overcome this limitation. In this paper, we introduce a triple point masking scheme, named TPM, which serves as a scalable framework for pre-training of masked autoencoders to achieve multi-mask learning for 3D point clouds. Specifically, we augment the baselines with two additional mask choices (i.e., medium mask and low mask) as our core insight is that the recovery process of an object can manifest in diverse ways. Previous high-masking schemes focus on capturing the global representation but lack the fine-grained recovery capability, so that the generated pre-trained weights tend to play a limited role in the fine-tuning process. With the support of the proposed TPM, available methods can exhibit more flexible and accurate completion capabilities, enabling the potential autoencoder in the pre-training stage to consider multiple representations of a single 3D object. In addition, an SVM-guided weight selection module is proposed to fill the encoder parameters for downstream networks with the optimal weight during the fine-tuning stage, maximizing linear accuracy and facilitating the acquisition of intricate representations for new objects. Extensive experiments show that the four baselines equipped with the proposed TPM achieve comprehensive performance improvements on various downstream tasks.
Authors: Pengcheng Shao, Tianyang Xu, Xuefeng Zhu, Xiaojun Wu, Josef Kittler
Abstract: Event-based bionic camera asynchronously captures dynamic scenes with high temporal resolution and high dynamic range, offering potential for the integration of events and RGB under conditions of illumination degradation and fast motion. Existing RGB-E tracking methods model event characteristics utilising attention mechanism of Transformer before integrating both modalities. Nevertheless, these methods involve aggregating the event stream into a single event frame, lacking the utilisation of the temporal information inherent in the event stream.Moreover, the traditional attention mechanism is well-suited for dense semantic features, while the attention mechanism for sparse event features require revolution. In this paper, we propose a dynamic event subframe splitting strategy to split the event stream into more fine-grained event clusters, aiming to capture spatio-temporal features that contain motion cues. Based on this, we design an event-based sparse attention mechanism to enhance the interaction of event features in temporal and spatial dimensions. The experimental results indicate that our method outperforms existing state-of-the-art methods on the FE240 and COESOT datasets, providing an effective processing manner for the event data.
Authors: Lingyi Hong, Jinglun Li, Xinyu Zhou, Shilin Yan, Pinxue Guo, Kaixun Jiang, Zhaoyu Chen, Shuyong Gao, Wei Zhang, Hong Lu, Wenqiang Zhang
Abstract: Transformer-based trackers have established a dominant role in the field of visual object tracking. While these trackers exhibit promising performance, their deployment on resource-constrained devices remains challenging due to inefficiencies. To improve the inference efficiency and reduce the computation cost, prior approaches have aimed to either design lightweight trackers or distill knowledge from larger teacher models into more compact student trackers. However, these solutions often sacrifice accuracy for speed. Thus, we propose a general model compression framework for efficient transformer object tracking, named CompressTracker, to reduce the size of a pre-trained tracking model into a lightweight tracker with minimal performance degradation. Our approach features a novel stage division strategy that segments the transformer layers of the teacher model into distinct stages, enabling the student model to emulate each corresponding teacher stage more effectively. Additionally, we also design a unique replacement training technique that involves randomly substituting specific stages in the student model with those from the teacher model, as opposed to training the student model in isolation. Replacement training enhances the student model's ability to replicate the teacher model's behavior. To further forcing student model to emulate teacher model, we incorporate prediction guidance and stage-wise feature mimicking to provide additional supervision during the teacher model's compression process. Our framework CompressTracker is structurally agnostic, making it compatible with any transformer architecture. We conduct a series of experiment to verify the effectiveness and generalizability of CompressTracker. Our CompressTracker-4 with 4 transformer layers, which is compressed from OSTrack, retains about 96% performance on LaSOT (66.1% AUC) while achieves 2.17x speed up.
Authors: Christina Zhang, Simran Motwani, Matthew Yu, Ji Hou, Felix Juefei-Xu, Sam Tsai, Peter Vajda, Zijian He, Jialiang Wang
Abstract: Latent diffusion models (LDMs) have made significant advancements in the field of image generation in recent years. One major advantage of LDMs is their ability to operate in a compressed latent space, allowing for more efficient training and deployment. However, despite these advantages, challenges with LDMs still remain. For example, it has been observed that LDMs often generate high-frequency details and complex compositions imperfectly. We hypothesize that one reason for these flaws is due to the fact that all pre- and post-training of LDMs are done in latent space, which is typically $8 \times 8$ lower spatial-resolution than the output images. To address this issue, we propose adding pixel-space supervision in the post-training process to better preserve high-frequency details. Experimentally, we show that adding a pixel-space objective significantly improves both supervised quality fine-tuning and preference-based post-training by a large margin on a state-of-the-art DiT transformer and U-Net diffusion models in both visual quality and visual flaw metrics, while maintaining the same text alignment quality.
Authors: Hongtao Huang, Xiaojun Chang, Lina Yao
Abstract: Diffusion models are cutting-edge generative models adept at producing diverse, high-quality images. Despite their effectiveness, these models often require significant computational resources owing to their numerous sequential denoising steps and the significant inference cost of each step. Recently, Neural Architecture Search (NAS) techniques have been employed to automatically search for faster generation processes. However, NAS for diffusion is inherently time-consuming as it requires estimating thousands of diffusion models to search for the optimal one. In this paper, we introduce Flexiffusion, a novel training-free NAS paradigm designed to accelerate diffusion models by concurrently optimizing generation steps and network structures. Specifically, we partition the generation process into isometric step segments, each sequentially composed of a full step, multiple partial steps, and several null steps. The full step computes all network blocks, while the partial step involves part of the blocks, and the null step entails no computation. Flexiffusion autonomously explores flexible step combinations for each segment, substantially reducing search costs and enabling greater acceleration compared to the state-of-the-art (SOTA) method for diffusion models. Our searched models reported speedup factors of $2.6\times$ and $1.5\times$ for the original LDM-4-G and the SOTA, respectively. The factors for Stable Diffusion V1.5 and the SOTA are $5.1\times$ and $2.0\times$. We also verified the performance of Flexiffusion on multiple datasets, and positive experiment results indicate that Flexiffusion can effectively reduce redundancy in diffusion models.
Authors: Shen Li, Jianqing Xu, Jiaying Wu, Miao Xiong, Ailin Deng, Jiazhen Ji, Yuge Huang, Wenjie Feng, Shouhong Ding, Bryan Hooi
Abstract: Synthetic face recognition (SFR) aims to generate synthetic face datasets that mimic the distribution of real face data, which allows for training face recognition models in a privacy-preserving manner. Despite the remarkable potential of diffusion models in image generation, current diffusion-based SFR models struggle with generalization to real-world faces. To address this limitation, we outline three key objectives for SFR: (1) promoting diversity across identities (inter-class diversity), (2) ensuring diversity within each identity by injecting various facial attributes (intra-class diversity), and (3) maintaining identity consistency within each identity group (intra-class identity preservation). Inspired by these goals, we introduce a diffusion-fueled SFR model termed $\text{ID}^3$. $\text{ID}^3$ employs an ID-preserving loss to generate diverse yet identity-consistent facial appearances. Theoretically, we show that minimizing this loss is equivalent to maximizing the lower bound of an adjusted conditional log-likelihood over ID-preserving data. This equivalence motivates an ID-preserving sampling algorithm, which operates over an adjusted gradient vector field, enabling the generation of fake face recognition datasets that approximate the distribution of real-world faces. Extensive experiments across five challenging benchmarks validate the advantages of $\text{ID}^3$.
Authors: Chengze Jiang, Junkai Wang, Minjing Dong, Jie Gui, Xinli Shi, Yuan Cao, Yuan Yan Tang, James Tin-Yau Kwok
Abstract: Adversarial training has achieved remarkable advancements in defending against adversarial attacks. Among them, fast adversarial training (FAT) is gaining attention for its ability to achieve competitive robustness with fewer computing resources. Existing FAT methods typically employ a uniform strategy that optimizes all training data equally without considering the influence of different examples, which leads to an imbalanced optimization. However, this imbalance remains unexplored in the field of FAT. In this paper, we conduct a comprehensive study of the imbalance issue in FAT and observe an obvious class disparity regarding their performances. This disparity could be embodied from a perspective of alignment between clean and robust accuracy. Based on the analysis, we mainly attribute the observed misalignment and disparity to the imbalanced optimization in FAT, which motivates us to optimize different training data adaptively to enhance robustness. Specifically, we take disparity and misalignment into consideration. First, we introduce self-knowledge guided regularization, which assigns differentiated regularization weights to each class based on its training state, alleviating class disparity. Additionally, we propose self-knowledge guided label relaxation, which adjusts label relaxation according to the training accuracy, alleviating the misalignment and improving robustness. By combining these methods, we formulate the Self-Knowledge Guided FAT (SKG-FAT), leveraging naturally generated knowledge during training to enhance the adversarial robustness without compromising training efficiency. Extensive experiments on four standard datasets demonstrate that the SKG-FAT improves the robustness and preserves competitive clean accuracy, outperforming the state-of-the-art methods.
Authors: Zhenyu Hu, Wanjie Sun
Abstract: Window-based transformers have demonstrated outstanding performance in super-resolution tasks due to their adaptive modeling capabilities through local self-attention (SA). However, they exhibit higher computational complexity and inference latency than convolutional neural networks. In this paper, we first identify that the adaptability of the Transformers is derived from their adaptive spatial aggregation and advanced structural design, while their high latency results from the computational costs and memory layout transformations associated with the local SA. To simulate this aggregation approach, we propose an effective convolution-based linear focal separable attention (FSA), allowing for long-range dynamic modeling with linear complexity. Additionally, we introduce an effective dual-branch structure combined with an ultra-lightweight information exchange module (IEM) to enhance the aggregation of information by the Token Mixer. Finally, with respect to the structure, we modify the existing spatial-gate-based feedforward neural networks by incorporating a self-gate mechanism to preserve high-dimensional channel information, enabling the modeling of more complex relationships. With these advancements, we construct a convolution-based Transformer framework named the linear adaptive mixer network (LAMNet). Extensive experiments demonstrate that LAMNet achieves better performance than existing SA-based Transformer methods while maintaining the computational efficiency of convolutional neural networks, which can achieve a \(3\times\) speedup of inference time. The code will be publicly available at: https://github.com/zononhzy/LAMNet.
Authors: Yuan Xun, Siyuan Liang, Xiaojun Jia, Xinwei Liu, Xiaochun Cao
Abstract: Pre-trained large models for multimodal contrastive learning, such as CLIP, have been widely recognized in the industry as highly susceptible to data-poisoned backdoor attacks. This poses significant risks to downstream model training. In response to such potential threats, finetuning offers a simpler and more efficient defense choice compared to retraining large models with augmented data. In the supervised learning domain, fine-tuning defense strategies can achieve excellent defense performance. However, in the unsupervised and semi-supervised domain, we find that when CLIP faces some complex attack techniques, the existing fine-tuning defense strategy, CleanCLIP, has some limitations on defense performance. The synonym substitution of its text-augmentation is insufficient to enhance the text feature space. To compensate for this weakness, we improve it by proposing a fine-grained \textbf{T}ext \textbf{A}lignment \textbf{C}leaner (TA-Cleaner) to cut off feature connections of backdoor triggers. We randomly select a few samples for positive and negative subtext generation at each epoch of CleanCLIP, and align the subtexts to the images to strengthen the text self-supervision. We evaluate the effectiveness of our TA-Cleaner against six attack algorithms and conduct comprehensive zero-shot classification tests on ImageNet1K. Our experimental results demonstrate that TA-Cleaner achieves state-of-the-art defensiveness among finetuning-based defense techniques. Even when faced with the novel attack technique BadCLIP, our TA-Cleaner outperforms CleanCLIP by reducing the ASR of Top-1 and Top-10 by 52.02\% and 63.88\%, respectively.
Authors: Amir Samadi, Konstantinos Koufos, Kurt Debattista, Mehrdad Dianati
Abstract: In this paper, we address the limitations of traditional teacher-student models, imitation learning, and behaviour cloning in the context of Autonomous/Automated Driving Systems (ADS), where these methods often struggle with incomplete coverage of real-world scenarios. To enhance the robustness of such models, we introduce the use of Counterfactual Explanations (CFEs) as a novel data augmentation technique for end-to-end ADS. CFEs, by generating training samples near decision boundaries through minimal input modifications, lead to a more comprehensive representation of expert driver strategies, particularly in safety-critical scenarios. This approach can therefore help improve the model's ability to handle rare and challenging driving events, such as anticipating darting out pedestrians, ultimately leading to safer and more trustworthy decision-making for ADS. Our experiments in the CARLA simulator demonstrate that CF-Driver outperforms the current state-of-the-art method, achieving a higher driving score and lower infraction rates. Specifically, CF-Driver attains a driving score of 84.2, surpassing the previous best model by 15.02 percentage points. These results highlight the effectiveness of incorporating CFEs in training end-to-end ADS. To foster further research, the CF-Driver code is made publicly available.
Authors: Jiahao Lyu, Minghua Zhao, Jing Hu, Xuewen Huang, Shuangli Du, Cheng Shi, Zhiyong Lv
Abstract: Video anomaly detection (VAD) often learns the distribution of normal samples and detects the anomaly through measuring significant deviations, but the undesired generalization may reconstruct a few anomalies thus suppressing the deviations. Meanwhile, most VADs cannot cope with cross-dataset validation for new target domains, and few-shot methods must laboriously rely on model-tuning from the target domain to complete domain adaptation. To address these problems, we propose a novel VAD method with a motion-guided memory module to achieve cross-dataset validation with zero-shot. First, we add Gaussian blur to the raw appearance images, thereby constructing the global pseudo-anomaly, which serves as the input to the network. Then, we propose multi-scale residual channel attention to deblur the pseudo-anomaly in normal samples. Next, memory items are obtained by recording the motion features in the training phase, which are used to retrieve the motion features from the raw information in the testing phase. Lastly, our method can ignore the blurred real anomaly through attention and rely on motion memory items to increase the normality gap between normal and abnormal motion. Extensive experiments on three benchmark datasets demonstrate the effectiveness of the proposed method. Compared with cross-domain methods, our method achieves competitive performance without adaptation during testing.
Authors: Taeyun Woo, Tae-Kyun Kim, Jinah Park
Abstract: Estimating the poses of both a hand and an object has become an important area of research due to the growing need for advanced vision computing. The primary challenge involves understanding and reconstructing how hands and objects interact, such as contact and physical plausibility. Existing approaches often adopt a graph neural network to incorporate spatial information of hand and object meshes. However, these approaches have not fully exploited the potential of graphs without modification of edges within and between hand- and object-graphs. We propose a graph-based refinement method that incorporates an interaction-aware graph-attention mechanism to account for hand-object interactions. Using edges, we establish connections among closely correlated nodes, both within individual graphs and across different graphs. Experiments demonstrate the effectiveness of our proposed method with notable improvements in the realm of physical plausibility.
Authors: Huixin Sun, Runqi Wang, Yanjing Li, Xianbin Cao, Xiaolong Jiang, Yao Hu, Baochang Zhang
Abstract: Large-scale pre-trained Vision-Language Models (VLMs) have gained prominence in various visual and multimodal tasks, yet the deployment of VLMs on downstream application platforms remains challenging due to their prohibitive requirements of training samples and computing resources. Fine-tuning and quantization of VLMs can substantially reduce the sample and computation costs, which are in urgent need. There are two prevailing paradigms in quantization, Quantization-Aware Training (QAT) can effectively quantize large-scale VLMs but incur a huge training cost, while low-bit Post-Training Quantization (PTQ) suffers from a notable performance drop. We propose a method that balances fine-tuning and quantization named ``Prompt for Quantization'' (P4Q), in which we design a lightweight architecture to leverage contrastive loss supervision to enhance the recognition performance of a PTQ model. Our method can effectively reduce the gap between image features and text features caused by low-bit quantization, based on learnable prompts to reorganize textual representations and a low-bit adapter to realign the distributions of image and text features. We also introduce a distillation loss based on cosine similarity predictions to distill the quantized model using a full-precision teacher. Extensive experimental results demonstrate that our P4Q method outperforms prior arts, even achieving comparable results to its full-precision counterparts. For instance, our 8-bit P4Q can theoretically compress the CLIP-ViT/B-32 by 4 $\times$ while achieving 66.94\% Top-1 accuracy, outperforming the learnable prompt fine-tuned full-precision model by 2.24\% with negligible additional parameters on the ImageNet dataset.
Authors: Tieyuan Chen, Huabin Liu, Tianyao He, Yihang Chen, Chaofan Gan, Xiao Ma, Cheng Zhong, Yang Zhang, Yingxue Wang, Hui Lin, Weiyao Lin
Abstract: Video causal reasoning aims to achieve a high-level understanding of video content from a causal perspective. However, current video reasoning tasks are limited in scope, primarily executed in a question-answering paradigm and focusing on short videos containing only a single event and simple causal relationships, lacking comprehensive and structured causality analysis for videos with multiple events. To fill this gap, we introduce a new task and dataset, Multi-Event Causal Discovery (MECD). It aims to uncover the causal relationships between events distributed chronologically across long videos. Given visual segments and textual descriptions of events, MECD requires identifying the causal associations between these events to derive a comprehensive, structured event-level video causal diagram explaining why and how the final result event occurred. To address MECD, we devise a novel framework inspired by the Granger Causality method, using an efficient mask-based event prediction model to perform an Event Granger Test, which estimates causality by comparing the predicted result event when premise events are masked versus unmasked. Furthermore, we integrate causal inference techniques such as front-door adjustment and counterfactual inference to address challenges in MECD like causality confounding and illusory causality. Experiments validate the effectiveness of our framework in providing causal relationships in multi-event videos, outperforming GPT-4o and VideoLLaVA by 5.7% and 4.1%, respectively.
Authors: Katja Ludwig, Julian Lorenz, Daniel Kienzle, Tuan Bui, Rainer Lienhart
Abstract: The basic body shape of a person does not change within a single video. However, most SOTA human mesh estimation (HME) models output a slightly different body shape for each video frame, which results in inconsistent body shapes for the same person. In contrast, we leverage anthropometric measurements like tailors are already obtaining from humans for centuries. We create a model called A2B that converts such anthropometric measurements to body shape parameters of human mesh models. Moreover, we find that finetuned SOTA 3D human pose estimation (HPE) models outperform HME models regarding the precision of the estimated keypoints. We show that applying inverse kinematics (IK) to the results of such a 3D HPE model and combining the resulting body pose with the A2B body shape leads to superior and consistent human meshes for challenging datasets like ASPset or fit3D, where we can lower the MPJPE by over 30 mm compared to SOTA HME models. Further, replacing HME models estimates of the body shape parameters with A2B model results not only increases the performance of these HME models, but also leads to consistent body shapes.
Authors: Huan Yang, Jiahui Chen, Chaofan Ding, Runhua Shi, Siyu Xiong, Qingqi Hong, Xiaoqi Mo, Xinhan Di
Abstract: Gestures are pivotal in enhancing co-speech communication. While recent works have mostly focused on point-level motion transformation or fully supervised motion representations through data-driven approaches, we explore the representation of gestures in co-speech, with a focus on self-supervised representation and pixel-level motion deviation, utilizing a diffusion model which incorporates latent motion features. Our approach leverages self-supervised deviation in latent representation to facilitate hand gestures generation, which are crucial for generating realistic gesture videos. Results of our first experiment demonstrate that our method enhances the quality of generated videos, with an improvement from 2.7 to 4.5% for FGD, DIV, and FVD, and 8.1% for PSNR, 2.5% for SSIM over the current state-of-the-art methods.
Authors: Ao Chang, Jiajun Zeng, Ruobing Huang, Dong Ni
Abstract: Convolutional neural networks have primarily led 3D medical image segmentation but may be limited by small receptive fields. Transformer models excel in capturing global relationships through self-attention but are challenged by high computational costs at high resolutions. Recently, Mamba, a state space model, has emerged as an effective approach for sequential modeling. Inspired by its success, we introduce a novel Mamba-based 3D medical image segmentation model called EM-Net. It not only efficiently captures attentive interaction between regions by integrating and selecting channels, but also effectively utilizes frequency domain to harmonize the learning of features across varying scales, while accelerating training speed. Comprehensive experiments on two challenging multi-organ datasets with other state-of-the-art (SOTA) algorithms show that our method exhibits better segmentation accuracy while requiring nearly half the parameter size of SOTA models and 2x faster training speed.
Authors: Suman Ghosh, Guillermo Gallego
Abstract: Stereopsis has widespread appeal in robotics as it is the predominant way by which living beings perceive depth to navigate our 3D world. Event cameras are novel bio-inspired sensors that detect per-pixel brightness changes asynchronously, with very high temporal resolution and high dynamic range, enabling machine perception in high-speed motion and broad illumination conditions. The high temporal precision also benefits stereo matching, making disparity (depth) estimation a popular research area for event cameras ever since its inception. Over the last 30 years, the field has evolved rapidly, from low-latency, low-power circuit design to current deep learning (DL) approaches driven by the computer vision community. The bibliography is vast and difficult to navigate for non-experts due its highly interdisciplinary nature. Past surveys have addressed distinct aspects of this topic, in the context of applications, or focusing only on a specific class of techniques, but have overlooked stereo datasets. This survey provides a comprehensive overview, covering both instantaneous stereo and long-term methods suitable for simultaneous localization and mapping (SLAM), along with theoretical and empirical comparisons. It is the first to extensively review DL methods as well as stereo datasets, even providing practical suggestions for creating new benchmarks to advance the field. The main advantages and challenges faced by event-based stereo depth estimation are also discussed. Despite significant progress, challenges remain in achieving optimal performance in not only accuracy but also efficiency, a cornerstone of event-based computing. We identify several gaps and propose future research directions. We hope this survey inspires future research in this area, by serving as an accessible entry point for newcomers, as well as a practical guide for seasoned researchers in the community.
Authors: Zheling Meng, Bo Peng, Xiaochuan Jin, Yue Jiang, Jing Dong, Wei Wang, Tieniu Tan
Abstract: Text-to-image diffusion models have been demonstrated with unsafe generation due to unfiltered large-scale training data, such as violent, sexual, and shocking images, necessitating the erasure of unsafe concepts. Most existing methods focus on modifying the generation probabilities conditioned on the texts containing unsafe descriptions. However, they fail to guarantee safe generation for unseen texts in the training phase, especially for the prompts from adversarial attacks. In this paper, we re-analyze the erasure task and point out that existing methods cannot guarantee the minimization of the total probabilities of unsafe generation. To tackle this problem, we propose Dark Miner. It entails a recurring three-stage process that comprises mining, verifying, and circumventing. It greedily mines embeddings with maximum generation probabilities of unsafe concepts and reduces unsafe generation more effectively. In the experiments, we evaluate its performance on two inappropriate concepts, two objects, and two styles. Compared with 6 previous state-of-the-art methods, our method achieves better erasure and defense results in most cases, especially under 4 state-of-the-art attacks, while preserving the model's native generation capability. Our code will be available on GitHub.
Authors: Weihao Yuan, Weichao Shen, Yisheng He, Yuan Dong, Xiaodong Gu, Zilong Dong, Liefeng Bo, Qixing Huang
Abstract: Motion generation from discrete quantization offers many advantages over continuous regression, but at the cost of inevitable approximation errors. Previous methods usually quantize the entire body pose into one code, which not only faces the difficulty in encoding all joints within one vector but also loses the spatial relationship between different joints. Differently, in this work we quantize each individual joint into one vector, which i) simplifies the quantization process as the complexity associated with a single joint is markedly lower than that of the entire pose; ii) maintains a spatial-temporal structure that preserves both the spatial relationships among joints and the temporal movement patterns; iii) yields a 2D token map, which enables the application of various 2D operations widely used in 2D images. Grounded in the 2D motion quantization, we build a spatial-temporal modeling framework, where 2D joint VQVAE, temporal-spatial 2D masking technique, and spatial-temporal 2D attention are proposed to take advantage of spatial-temporal signals among the 2D tokens. Extensive experiments demonstrate that our method significantly outperforms previous methods across different datasets, with a $26.6\%$ decrease of FID on HumanML3D and a $29.9\%$ decrease on KIT-ML.
Authors: Dimitrios Kollias, Chunchang Shao, Odysseus Kaloidas, Ioannis Patras
Abstract: In this paper, we introduce Behavior4All, a comprehensive, open-source toolkit for in-the-wild facial behavior analysis, integrating Face Localization, Valence-Arousal Estimation, Basic Expression Recognition and Action Unit Detection, all within a single framework. Available in both CPU-only and GPU-accelerated versions, Behavior4All leverages 12 large-scale, in-the-wild datasets consisting of over 5 million images from diverse demographic groups. It introduces a novel framework that leverages distribution matching and label co-annotation to address tasks with non-overlapping annotations, encoding prior knowledge of their relatedness. In the largest study of its kind, Behavior4All outperforms both state-of-the-art and toolkits in overall performance as well as fairness across all databases and tasks. It also demonstrates superior generalizability on unseen databases and on compound expression recognition. Finally, Behavior4All is way times faster than other toolkits.
Authors: Seraj Ghasemi, Hamed Hosseini, MohammadHossein Koosheshi, Mehdi Tale Masouleh, Ahmad Kalhor
Abstract: With robots increasingly collaborating with humans in everyday tasks, it is important to take steps toward robotic systems capable of understanding the environment. This work focuses on scene understanding to detect pick and place tasks given initial and final images from the scene. To this end, a dataset is collected for object detection and pick and place task detection. A YOLOv5 network is subsequently trained to detect the objects in the initial and final scenes. Given the detected objects and their bounding boxes, two methods are proposed to detect the pick and place tasks which transform the initial scene into the final scene. A geometric method is proposed which tracks objects' movements in the two scenes and works based on the intersection of the bounding boxes which moved within scenes. Contrarily, the CNN-based method utilizes a Convolutional Neural Network to classify objects with intersected bounding boxes into 5 classes, showing the spatial relationship between the involved objects. The performed pick and place tasks are then derived from analyzing the experiments with both scenes. Results show that the CNN-based method, using a VGG16 backbone, outscores the geometric method by roughly 12 percentage points in certain scenarios, with an overall success rate of 84.3%.
Authors: Shiqi Sun, Yantao Lu, Ning Liu, Bo Jiang, JinChao Chen, Ying Zhang
Abstract: Camera-LiDAR fusion models significantly enhance perception performance in autonomous driving. The fusion mechanism leverages the strengths of each modality while minimizing their weaknesses. Moreover, in practice, camera-LiDAR fusion models utilize pre-trained backbones for efficient training. However, we argue that directly loading single-modal pre-trained camera and LiDAR backbones into camera-LiDAR fusion models introduces similar feature redundancy across modalities due to the nature of the fusion mechanism. Unfortunately, existing pruning methods are developed explicitly for single-modal models, and thus, they struggle to effectively identify these specific redundant parameters in camera-LiDAR fusion models. In this paper, to address the issue above on camera-LiDAR fusion models, we propose a novelty pruning framework Alternative Modality Masking Pruning (AlterMOMA), which employs alternative masking on each modality and identifies the redundant parameters. Specifically, when one modality parameters are masked (deactivated), the absence of features from the masked backbone compels the model to reactivate previous redundant features of the other modality backbone. Therefore, these redundant features and relevant redundant parameters can be identified via the reactivation process. The redundant parameters can be pruned by our proposed importance score evaluation function, Alternative Evaluation (AlterEva), which is based on the observation of the loss changes when certain modality parameters are activated and deactivated. Extensive experiments on the nuScene and KITTI datasets encompassing diverse tasks, baseline models, and pruning algorithms showcase that AlterMOMA outperforms existing pruning methods, attaining state-of-the-art performance.
Authors: Qi Zhang, He Wang, Ru Li, Wenbin Li
Abstract: Recent advancements in Simultaneous Localization and Mapping (SLAM) have increasingly highlighted the robustness of LiDAR-based techniques. At the same time, Neural Radiance Fields (NeRF) have introduced new possibilities for 3D scene reconstruction, exemplified by SLAM systems. Among these, NeRF-LOAM has shown notable performance in NeRF-based SLAM applications. However, despite its strengths, these systems often encounter difficulties in dynamic outdoor environments due to their inherent static assumptions. To address these limitations, this paper proposes a novel method designed to improve reconstruction in highly dynamic outdoor scenes. Based on NeRF-LOAM, the proposed approach consists of two primary components. First, we separate the scene into static background and dynamic foreground. By identifying and excluding dynamic elements from the mapping process, this segmentation enables the creation of a dense 3D map that accurately represents the static background only. The second component extends the octree structure to support multi-resolution representation. This extension not only enhances reconstruction quality but also aids in the removal of dynamic objects identified by the first module. Additionally, Fourier feature encoding is applied to the sampled points, capturing high-frequency information and leading to more complete reconstruction results. Evaluations on various datasets demonstrate that our method achieves more competitive results compared to current state-of-the-art approaches.
Authors: Jinghao Zhang, Wen Qian, Hao Luo, Fan Wang, Feng Zhao
Abstract: Diffusion models have made compelling progress on facilitating high-throughput daily production. Nevertheless, the appealing customized requirements are remain suffered from instance-level finetuning for authentic fidelity. Prior zero-shot customization works achieve the semantic consistence through the condensed injection of identity features, while addressing detailed low-level signatures through complex model configurations and subject-specific fabrications, which significantly break the statistical coherence within the overall system and limit the applicability across various scenarios. To facilitate the generic signature concentration with rectified efficiency, we present \textbf{AnyLogo}, a zero-shot region customizer with remarkable detail consistency, building upon the symbiotic diffusion system with eliminated cumbersome designs. Streamlined as vanilla image generation, we discern that the rigorous signature extraction and creative content generation are promisingly compatible and can be systematically recycled within a single denoising model. In place of the external configurations, the gemini status of the denoising model promote the reinforced subject transmission efficiency and disentangled semantic-signature space with continuous signature decoration. Moreover, the sparse recycling paradigm is adopted to prevent the duplicated risk with compressed transmission quota for diversified signature stimulation. Extensive experiments on constructed logo-level benchmarks demonstrate the effectiveness and practicability of our methods.
Authors: Chihiro Noguchi, Shun Fukuda, Shoichiro Mihara, Masao Yamanaka
Abstract: Scene text recognition in low-resource languages frequently faces challenges due to the limited availability of training datasets derived from real-world scenes. This study proposes a novel approach that generates text images in low-resource languages by emulating the style of real text images from high-resource languages. Our approach utilizes a diffusion model that is conditioned on binary states: ``synthetic'' and ``real.'' The training of this model involves dual translation tasks, where it transforms plain text images into either synthetic or real text images, based on the binary states. This approach not only effectively differentiates between the two domains but also facilitates the model's explicit recognition of characters in the target language. Furthermore, to enhance the accuracy and variety of generated text images, we introduce two guidance techniques: Fidelity-Diversity Balancing Guidance and Fidelity Enhancement Guidance. Our experimental results demonstrate that the text images generated by our proposed framework can significantly improve the performance of scene text recognition models for low-resource languages.
Authors: Evangelia Christodoulou, Annika Reinke, Rola Houhou, Piotr Kalinowski, Selen Erkan, Carole H. Sudre, Ninon Burgos, Sofi\`ene Boutaj, Sophie Loizillon, Ma\"elys Solal, Nicola Rieke, Veronika Cheplygina, Michela Antonelli, Leon D. Mayer, Minu D. Tizabi, M. Jorge Cardoso, Amber Simpson, Paul F. J\"ager, Annette Kopp-Schneider, Ga\"el Varoquaux, Olivier Colliot, Lena Maier-Hein
Abstract: Medical imaging is spearheading the AI transformation of healthcare. Performance reporting is key to determine which methods should be translated into clinical practice. Frequently, broad conclusions are simply derived from mean performance values. In this paper, we argue that this common practice is often a misleading simplification as it ignores performance variability. Our contribution is threefold. (1) Analyzing all MICCAI segmentation papers (n = 221) published in 2023, we first observe that more than 50\% of papers do not assess performance variability at all. Moreover, only one (0.5\%) paper reported confidence intervals (CIs) for model performance. (2) To address the reporting bottleneck, we show that the unreported standard deviation (SD) in segmentation papers can be approximated by a second-order polynomial function of the mean Dice similarity coefficient (DSC). Based on external validation data from 56 previous MICCAI challenges, we demonstrate that this approximation can accurately reconstruct the CI of a method using information provided in publications. (3) Finally, we reconstructed 95\% CIs around the mean DSC of MICCAI 2023 segmentation papers. The median CI width was 0.03 which is three times larger than the median performance gap between the first and second ranked method. For more than 60\% of papers, the mean performance of the second-ranked method was within the CI of the first-ranked method. We conclude that current publications typically do not provide sufficient evidence to support which models could potentially be translated into clinical practice.
Authors: Valentin Koch, Sabine Bauer, Valerio Luppberger, Michael Joner, Heribert Schunkert, Julia A. Schnabel, Moritz von Scheidt, Carsten Marr
Abstract: Background: The integration of multi-stain histopathology images through deep learning poses a significant challenge in digital histopathology. Current multi-modal approaches struggle with data heterogeneity and missing data. This study aims to overcome these limitations by developing a novel transformer model for multi-stain integration that can handle missing data during training as well as inference. Methods: We propose UNICORN (UNiversal modality Integration Network for CORonary classificatioN) a multi-modal transformer capable of processing multi-stain histopathology for atherosclerosis severity class prediction. The architecture comprises a two-stage, end-to-end trainable model with specialized modules utilizing transformer self-attention blocks. The initial stage employs domain-specific expert modules to extract features from each modality. In the subsequent stage, an aggregation expert module integrates these features by learning the interactions between the different data modalities. Results: Evaluation was performed using a multi-class dataset of atherosclerotic lesions from the Munich Cardiovascular Studies Biobank (MISSION), using over 4,000 paired multi-stain whole slide images (WSIs) from 170 deceased individuals on 7 prespecified segments of the coronary tree, each stained according to four histopathological protocols. UNICORN achieved a classification accuracy of 0.67, outperforming other state-of-the-art models. The model effectively identifies relevant tissue phenotypes across stainings and implicitly models disease progression. Conclusion: Our proposed multi-modal transformer model addresses key challenges in medical data analysis, including data heterogeneity and missing modalities. Explainability and the model's effectiveness in predicting atherosclerosis progression underscores its potential for broader applications in medical research.
Authors: Raja Kumar, Raghav Singhal, Pranamya Kulkarni, Deval Mehta, Kshitij Jadhav
Abstract: Deep multimodal learning has shown remarkable success by leveraging contrastive learning to capture explicit one-to-one relations across modalities. However, real-world data often exhibits shared relations beyond simple pairwise associations. We propose M3CoL, a Multimodal Mixup Contrastive Learning approach to capture nuanced shared relations inherent in multimodal data. Our key contribution is a Mixup-based contrastive loss that learns robust representations by aligning mixed samples from one modality with their corresponding samples from other modalities thereby capturing shared relations between them. For multimodal classification tasks, we introduce a framework that integrates a fusion module with unimodal prediction modules for auxiliary supervision during training, complemented by our proposed Mixup-based contrastive loss. Through extensive experiments on diverse datasets (N24News, ROSMAP, BRCA, and Food-101), we demonstrate that M3CoL effectively captures shared multimodal relations and generalizes across domains. It outperforms state-of-the-art methods on N24News, ROSMAP, and BRCA, while achieving comparable performance on Food-101. Our work highlights the significance of learning shared relations for robust multimodal learning, opening up promising avenues for future research.
Authors: Qinpeng Cui, Yixuan Liu, Xinyi Zhang, Qiqi Bao, Zhongdao Wang, Qingmin Liao, Li Wang, Tian Lu, Emad Barsoum
Abstract: Diffusion-based image super-resolution (SR) models have attracted substantial interest due to their powerful image restoration capabilities. However, prevailing diffusion models often struggle to strike an optimal balance between efficiency and performance. Typically, they either neglect to exploit the potential of existing extensive pretrained models, limiting their generative capacity, or they necessitate a dozens of forward passes starting from random noises, compromising inference efficiency. In this paper, we present DoSSR, a Domain Shift diffusion-based SR model that capitalizes on the generative powers of pretrained diffusion models while significantly enhancing efficiency by initiating the diffusion process with low-resolution (LR) images. At the core of our approach is a domain shift equation that integrates seamlessly with existing diffusion models. This integration not only improves the use of diffusion prior but also boosts inference efficiency. Moreover, we advance our method by transitioning the discrete shift process to a continuous formulation, termed as DoS-SDEs. This advancement leads to the fast and customized solvers that further enhance sampling efficiency. Empirical results demonstrate that our proposed method achieves state-of-the-art performance on synthetic and real-world datasets, while notably requiring only 5 sampling steps. Compared to previous diffusion prior based methods, our approach achieves a remarkable speedup of 5-7 times, demonstrating its superior efficiency. Code: https://github.com/QinpengCui/DoSSR.
Authors: Xinya Shu, Yu Li, Dongwei Ren, Xiaohe Wu, Jin Li, Wangmeng Zuo
Abstract: For single image defocus deblurring, acquiring well-aligned training pairs (or training triplets), i.e., a defocus blurry image, an all-in-focus sharp image (and a defocus blur map), is an intricate task for the development of deblurring models. Existing image defocus deblurring methods typically rely on training data collected by specialized imaging equipment, presupposing that these pairs or triplets are perfectly aligned. However, in practical scenarios involving the collection of real-world data, direct acquisition of training triplets is infeasible, and training pairs inevitably encounter spatial misalignment issues. In this work, we introduce a reblurring-guided learning framework for single image defocus deblurring, enabling the learning of a deblurring network even with misaligned training pairs. Specifically, we first propose a baseline defocus deblurring network that utilizes spatially varying defocus blur map as degradation prior to enhance the deblurring performance. Then, to effectively learn the baseline defocus deblurring network with misaligned training pairs, our reblurring module ensures spatial consistency between the deblurred image, the reblurred image and the input blurry image by reconstructing spatially variant isotropic blur kernels. Moreover, the spatially variant blur derived from the reblurring module can serve as pseudo supervision for defocus blur map during training, interestingly transforming training pairs into training triplets. Additionally, we have collected a new dataset specifically for single image defocus deblurring (SDD) with typical misalignments, which not only substantiates our proposed method but also serves as a benchmark for future research.
Authors: Ge Wu, Xin Zhang, Zheng Li, Zhaowei Chen, Jiajun Liang, Jian Yang, Xiang Li
Abstract: Prompt learning has surfaced as an effective approach to enhance the performance of Vision-Language Models (VLMs) like CLIP when applied to downstream tasks. However, current learnable prompt tokens are primarily used for the single phase of adapting to tasks (i.e., adapting prompt), easily leading to overfitting risks. In this work, we propose a novel Cascade Prompt Learning CasPL framework to enable prompt learning to serve both generic and specific expertise (i.e., boosting and adapting prompt) simultaneously. Specifically, CasPL is a new learning paradigm comprising two distinct phases of learnable prompts: the first boosting prompt is crafted to extract domain-general knowledge from a senior larger CLIP teacher model by aligning their predicted logits using extensive unlabeled domain images. The second adapting prompt is then cascaded with the frozen first set to fine-tune the downstream tasks, following the approaches employed in prior research. In this manner, CasPL can effectively capture both domain-general and task-specific representations into explicitly different gradual groups of prompts, thus potentially alleviating overfitting issues in the target domain. It's worth noting that CasPL serves as a plug-and-play module that can seamlessly integrate into any existing prompt learning approach. CasPL achieves a significantly better balance between performance and inference speed, which is especially beneficial for deploying smaller VLM models in resource-constrained environments. Compared to the previous state-of-the-art method PromptSRC, CasPL shows an average improvement of 1.85% for base classes, 3.44% for novel classes, and 2.72% for the harmonic mean over 11 image classification datasets. Code is publicly available at: https://github.com/megvii-research/CasPL.
Authors: Yuchen Guan, Runxi Cheng, Kang Liu, Chun Yuan
Abstract: Knowledge distillation typically employs the Kullback-Leibler (KL) divergence to constrain the student model's output to match the soft labels provided by the teacher model exactly. However, sometimes the optimization direction of the KL divergence loss is not always aligned with the task loss, where a smaller KL divergence could lead to erroneous predictions that diverge from the soft labels. This limitation often results in suboptimal optimization for the student. Moreover, even under temperature scaling, the KL divergence loss function tends to overly focus on the larger-valued channels in the logits, disregarding the rich inter-class information provided by the multitude of smaller-valued channels. This hard constraint proves too challenging for lightweight students, hindering further knowledge distillation. To address this issue, we propose a plug-and-play ranking loss based on Kendall's $\tau$ coefficient, called Rank-Kendall Knowledge Distillation (RKKD). RKKD balances the attention to smaller-valued channels by constraining the order of channel values in student logits, providing more inter-class relational information. The rank constraint on the top-valued channels helps avoid suboptimal traps during optimization. We also discuss different differentiable forms of Kendall's $\tau$ coefficient and demonstrate that the proposed ranking loss function shares a consistent optimization objective with the KL divergence. Extensive experiments on the CIFAR-100 and ImageNet datasets show that our RKKD can enhance the performance of various knowledge distillation baselines and offer broad improvements across multiple teacher-student architecture combinations.
Authors: Chaobing Zheng, Shiqian Wu, Zhenggguo Li
Abstract: Unsupervised learning based multi-scale exposure fusion (ULMEF) is efficient for fusing differently exposed low dynamic range (LDR) images into a higher quality LDR image for a high dynamic range (HDR) scene. Unlike supervised learning, loss functions play a crucial role in the ULMEF. In this paper, novel loss functions are proposed for the ULMEF and they are defined by using all the images to be fused and other differently exposed images from the same HDR scene. The proposed loss functions can guide the proposed ULMEF to learn more reliable information from the HDR scene than existing loss functions which are defined by only using the set of images to be fused. As such, the quality of the fused image is significantly improved. The proposed ULMEF also adopts a multi-scale strategy that includes a multi-scale attention module to effectively preserve the scene depth and local contrast in the fused image. Meanwhile, the proposed ULMEF can be adopted to achieve exposure interpolation and exposure extrapolation. Extensive experiments show that the proposed ULMEF algorithm outperforms state-of-the-art exposure fusion algorithms.
Authors: Aurel Pjetri (Verizon Connect Research, Florence, Italy, Department of Information Engineering, University of Florence, Florence, Italy), Stefano Caprasecca (Verizon Connect Research, Florence, Italy), Leonardo Taccari (Verizon Connect Research, Florence, Italy), Matteo Simoncini (Verizon Connect Research, Florence, Italy), Henrique Pi\~neiro Monteagudo (Verizon Connect Research, Florence, Italy, University of Bologna, Bologna, Italy), Walter Wallace (Verizon Connect Research, Florence, Italy), Douglas Coimbra de Andrade (SENAI Institute of Innovation, Rio de Janeiro, Brazil), Francesco Sambo (Verizon Connect Research, Florence, Italy), Andrew David Bagdanov (Verizon Connect Research, Florence, Italy)
Abstract: Monocular depth estimation is a critical task for autonomous driving and many other computer vision applications. While significant progress has been made in this field, the effects of viewpoint shifts on depth estimation models remain largely underexplored. This paper introduces a novel dataset and evaluation methodology to quantify the impact of different camera positions and orientations on monocular depth estimation performance. We propose a ground truth strategy based on homography estimation and object detection, eliminating the need for expensive lidar sensors. We collect a diverse dataset of road scenes from multiple viewpoints and use it to assess the robustness of a modern depth estimation model to geometric shifts. After assessing the validity of our strategy on a public dataset, we provide valuable insights into the limitations of current models and highlight the importance of considering viewpoint variations in real-world applications.
Authors: Jiaqi Li, Yiran Wang, Jinghong Zheng, Zihao Huang, Ke Xian, Zhiguo Cao, Jianming Zhang
Abstract: Depth refinement aims to infer high-resolution depth with fine-grained edges and details, refining low-resolution results of depth estimation models. The prevailing methods adopt tile-based manners by merging numerous patches, which lacks efficiency and produces inconsistency. Besides, prior arts suffer from fuzzy depth boundaries and limited generalizability. Analyzing the fundamental reasons for these limitations, we model depth refinement as a noisy Poisson fusion problem with local inconsistency and edge deformation noises. We propose the Self-distilled Depth Refinement (SDDR) framework to enforce robustness against the noises, which mainly consists of depth edge representation and edge-based guidance. With noisy depth predictions as input, SDDR generates low-noise depth edge representations as pseudo-labels by coarse-to-fine self-distillation. Edge-based guidance with edge-guided gradient loss and edge-based fusion loss serves as the optimization objective equivalent to Poisson fusion. When depth maps are better refined, the labels also become more noise-free. Our model can acquire strong robustness to the noises, achieving significant improvements in accuracy, edge quality, efficiency, and generalizability on five different benchmarks. Moreover, directly training another model with edge labels produced by SDDR brings improvements, suggesting that our method could help with training robust refinement models in future works.
Authors: Andrea Toaiari, Vittorio Murino, Marco Cristani, Cigdem Beyan
Abstract: Gaze Target Detection (GTD), i.e., determining where a person is looking within a scene from an external viewpoint, is a challenging task, particularly in 3D space. Existing approaches heavily rely on analyzing the person's appearance, primarily focusing on their face to predict the gaze target. This paper presents a novel approach to tackle this problem by utilizing the person's upper-body pose and available depth maps to extract a 3D gaze direction and employing a multi-stage or an end-to-end pipeline to predict the gazed target. When predicted accurately, the human body pose can provide valuable information about the head pose, which is a good approximation of the gaze direction, as well as the position of the arms and hands, which are linked to the activity the person is performing and the objects they are likely focusing on. Consequently, in addition to performing gaze estimation in 3D, we are also able to perform GTD simultaneously. We demonstrate state-of-the-art results on the most comprehensive publicly accessible 3D gaze target detection dataset without requiring images of the person's face, thus promoting privacy preservation in various application contexts. The code is available at https://github.com/intelligolabs/privacy-gtd-3D.
Authors: Xuezhi Xiang, Yao Wang, Lei Zhang, Denis Ombati, Himaloy Himu, Xiantong Zhen
Abstract: Self-supervised monocular depth estimation has emerged as a promising approach since it does not rely on labeled training data. Most methods combine convolution and Transformer to model long-distance dependencies to estimate depth accurately. However, Transformer treats 2D image features as 1D sequences, and positional encoding somewhat mitigates the loss of spatial information between different feature blocks, tending to overlook channel features, which limit the performance of depth estimation. In this paper, we propose a self-supervised monocular depth estimation network to get finer details. Specifically, we propose a decoder based on large kernel attention, which can model long-distance dependencies without compromising the two-dimension structure of features while maintaining feature channel adaptivity. In addition, we introduce a up-sampling module to accurately recover the fine details in the depth map. Our method achieves competitive results on the KITTI dataset.
Authors: Xuezhi Xiang, Zhushan Ma, Lei Zhang, Denis Ombati, Himaloy Himu, Xiantong Zhen
Abstract: With the rapid development of intelligent transportation systems and the popularity of smart city infrastructure, Vehicle Re-ID technology has become an important research field. The vehicle Re-ID task faces an important challenge, which is the high similarity between different vehicles. Existing methods use additional detection or segmentation models to extract differentiated local features. However, these methods either rely on additional annotations or greatly increase the computational cost. Using attention mechanism to capture global and local features is crucial to solve the challenge of high similarity between classes in vehicle Re-ID tasks. In this paper, we propose LKA-ReID with large kernel attention. Specifically, the large kernel attention (LKA) utilizes the advantages of self-attention and also benefits from the advantages of convolution, which can extract the global and local features of the vehicle more comprehensively. We also introduce hybrid channel attention (HCA) combines channel attention with spatial information, so that the model can better focus on channels and feature regions, and ignore background and other disturbing information. Experiments on VeRi-776 dataset demonstrated the effectiveness of LKA-ReID, with mAP reaches 86.65% and Rank-1 reaches 98.03%.
Authors: Dmytro Kotovenko, Olga Grebenkova, Nikolaos Sarafianos, Avinash Paliwal, Pingchuan Ma, Omid Poursaeed, Sreyas Mohan, Yuchen Fan, Yilei Li, Rakesh Ranjan, Bj\"orn Ommer
Abstract: While style transfer techniques have been well-developed for 2D image stylization, the extension of these methods to 3D scenes remains relatively unexplored. Existing approaches demonstrate proficiency in transferring colors and textures but often struggle with replicating the geometry of the scenes. In our work, we leverage an explicit Gaussian Splatting (GS) representation and directly match the distributions of Gaussians between style and content scenes using the Earth Mover's Distance (EMD). By employing the entropy-regularized Wasserstein-2 distance, we ensure that the transformation maintains spatial smoothness. Additionally, we decompose the scene stylization problem into smaller chunks to enhance efficiency. This paradigm shift reframes stylization from a pure generative process driven by latent space losses to an explicit matching of distributions between two Gaussian representations. Our method achieves high-resolution 3D stylization by faithfully transferring details from 3D style scenes onto the content scene. Furthermore, WaSt-3D consistently delivers results across diverse content and style scenes without necessitating any training, as it relies solely on optimization-based techniques. See our project page for additional results and source code: $\href{https://compvis.github.io/wast3d/}{https://compvis.github.io/wast3d/}$.
URLs: https://compvis.github.io/wast3d/, https://compvis.github.io/wast3d/
Authors: Qihan Huang, Siming Fu, Jinlong Liu, Hao Jiang, Yipeng Yu, Jie Song
Abstract: Personalized text-to-image generation methods can generate customized images based on the reference images, which have garnered wide research interest. Recent methods propose a finetuning-free approach with a decoupled cross-attention mechanism to generate personalized images requiring no test-time finetuning. However, when multiple reference images are provided, the current decoupled cross-attention mechanism encounters the object confusion problem and fails to map each reference image to its corresponding object, thereby seriously limiting its scope of application. To address the object confusion problem, in this work we investigate the relevance of different positions of the latent image features to the target object in diffusion model, and accordingly propose a weighted-merge method to merge multiple reference image features into the corresponding objects. Next, we integrate this weighted-merge method into existing pre-trained models and continue to train the model on a multi-object dataset constructed from the open-sourced SA-1B dataset. To mitigate object confusion and reduce training costs, we propose an object quality score to estimate the image quality for the selection of high-quality training samples. Furthermore, our weighted-merge training framework can be employed on single-object generation when a single object has multiple reference images. The experiments verify that our method achieves superior performance to the state-of-the-arts on the Concept101 dataset and DreamBooth dataset of multi-object personalized image generation, and remarkably improves the performance on single-object personalized image generation. Our code is available at https://github.com/hqhQAQ/MIP-Adapter.
Authors: Ilya Chugunov, Amogh Joshi, Kiran Murthy, Francois Bleibel, Felix Heide
Abstract: Challenging to capture, and challenging to display on a cellphone screen, the panorama paradoxically remains both a staple and underused feature of modern mobile camera applications. In this work we address both of these challenges with a spherical neural light field model for implicit panoramic image stitching and re-rendering; able to accommodate for depth parallax, view-dependent lighting, and local scene motion and color changes during capture. Fit during test-time to an arbitrary path panoramic video capture -- vertical, horizontal, random-walk -- these neural light spheres jointly estimate the camera path and a high-resolution scene reconstruction to produce novel wide field-of-view projections of the environment. Our single-layer model avoids expensive volumetric sampling, and decomposes the scene into compact view-dependent ray offset and color components, with a total model size of 80 MB per scene, and real-time (50 FPS) rendering at 1080p resolution. We demonstrate improved reconstruction quality over traditional image stitching and radiance field methods, with significantly higher tolerance to scene motion and non-ideal capture settings.
Authors: Filippo Bartolucci, Iacopo Masi, Giuseppe Lisanti
Abstract: Image manipulation detection and localization have received considerable attention from the research community given the blooming of Generative Models (GMs). Detection methods that follow a passive approach may overfit to specific GMs, limiting their application in real-world scenarios, due to the growing diversity of generative models. Recently, approaches based on a proactive framework have shown the possibility of dealing with this limitation. However, these methods suffer from two main limitations, which raises concerns about potential vulnerabilities: i) the manipulation detector is not robust to noise and hence can be easily fooled; ii) the fact that they rely on fixed perturbations for image protection offers a predictable exploit for malicious attackers, enabling them to reverse-engineer and evade detection. To overcome this issue we propose PADL, a new solution able to generate image-specific perturbations using a symmetric scheme of encoding and decoding based on cross-attention, which drastically reduces the possibility of reverse engineering, even when evaluated with adaptive attack [31]. Additionally, PADL is able to pinpoint manipulated areas, facilitating the identification of specific regions that have undergone alterations, and has more generalization power than prior art on held-out generative models. Indeed, although being trained only on an attribute manipulation GAN model [15], our method generalizes to a range of unseen models with diverse architectural designs, such as StarGANv2, BlendGAN, DiffAE, StableDiffusion and StableDiffusionXL. Additionally, we introduce a novel evaluation protocol, which offers a fair evaluation of localisation performance in function of detection accuracy and better captures real-world scenarios.
Authors: Xinpeng Yin, Wenming Cao
Abstract: In self-supervised skeleton-based action recognition, the mask reconstruction paradigm is gaining interest in enhancing model refinement and robustness through effective masking. However, previous works primarily relied on a single masking criterion, resulting in the model overfitting specific features and overlooking other effective information. In this paper, we introduce a hierarchy and attention guided cross-masking framework (HA-CM) that applies masking to skeleton sequences from both spatial and temporal perspectives. Specifically, in spatial graphs, we utilize hyperbolic space to maintain joint distinctions and effectively preserve the hierarchical structure of high-dimensional skeletons, employing joint hierarchy as the masking criterion. In temporal flows, we substitute traditional distance metrics with the global attention of joints for masking, addressing the convergence of distances in high-dimensional space and the lack of a global perspective. Additionally, we incorporate cross-contrast loss based on the cross-masking framework into the loss function to enhance the model's learning of instance-level features. HA-CM shows efficiency and universality on three public large-scale datasets, NTU-60, NTU-120, and PKU-MMD. The source code of our HA-CM is available at https://github.com/YinxPeng/HA-CM-main.
Authors: Linye Lyu, Jiawei Zhou, Daojing He, Yu Li
Abstract: Prior works on physical adversarial camouflage against vehicle detectors mainly focus on the effectiveness and robustness of the attack. The current most successful methods optimize 3D vehicle texture at a pixel level. However, this results in conspicuous and attention-grabbing patterns in the generated camouflage, which humans can easily identify. To address this issue, we propose a Customizable and Natural Camouflage Attack (CNCA) method by leveraging an off-the-shelf pre-trained diffusion model. By sampling the optimal texture image from the diffusion model with a user-specific text prompt, our method can generate natural and customizable adversarial camouflage while maintaining high attack performance. With extensive experiments on the digital and physical worlds and user studies, the results demonstrate that our proposed method can generate significantly more natural-looking camouflage than the state-of-the-art baselines while achieving competitive attack performance. Our code is available at \href{https://anonymous.4open.science/r/CNCA-1D54}{https://anonymous.4open.science/r/CNCA-1D54}
URLs: https://anonymous.4open.science/r/CNCA-1D54, https://anonymous.4open.science/r/CNCA-1D54
Authors: Yunpeng Gong, Qingyuan Zeng, Dejun Xu, Zhenzhong Wang, Min Jiang
Abstract: In recent years, despite significant advancements in adversarial attack research, the security challenges in cross-modal scenarios, such as the transferability of adversarial attacks between infrared, thermal, and RGB images, have been overlooked. These heterogeneous image modalities collected by different hardware devices are widely prevalent in practical applications, and the substantial differences between modalities pose significant challenges to attack transferability. In this work, we explore a novel cross-modal adversarial attack strategy, termed multiform attack. We propose a dual-layer optimization framework based on gradient-evolution, facilitating efficient perturbation transfer between modalities. In the first layer of optimization, the framework utilizes image gradients to learn universal perturbations within each modality and employs evolutionary algorithms to search for shared perturbations with transferability across different modalities through secondary optimization. Through extensive testing on multiple heterogeneous datasets, we demonstrate the superiority and robustness of Multiform Attack compared to existing techniques. This work not only enhances the transferability of cross-modal adversarial attacks but also provides a new perspective for understanding security vulnerabilities in cross-modal systems.
Authors: Janek Haberer, Ali Hojjat, Olaf Landsiedel
Abstract: The architecture of Vision Transformers (ViTs), particularly the Multi-head Attention (MHA) mechanism, imposes substantial hardware demands. Deploying ViTs on devices with varying constraints, such as mobile phones, requires multiple models of different sizes. However, this approach has limitations, such as training and storing each required model separately. This paper introduces HydraViT, a novel approach that addresses these limitations by stacking attention heads to achieve a scalable ViT. By repeatedly changing the size of the embedded dimensions throughout each layer and their corresponding number of attention heads in MHA during training, HydraViT induces multiple subnetworks. Thereby, HydraViT achieves adaptability across a wide spectrum of hardware environments while maintaining performance. Our experimental results demonstrate the efficacy of HydraViT in achieving a scalable ViT with up to 10 subnetworks, covering a wide range of resource constraints. HydraViT achieves up to 5 p.p. more accuracy with the same GMACs and up to 7 p.p. more accuracy with the same throughput on ImageNet-1K compared to the baselines, making it an effective solution for scenarios where hardware availability is diverse or varies over time. Source code available at https://github.com/ds-kiel/HydraViT.
Authors: Yichen Shen, Yijin Li, Shuo Chen, Guanglin Li, Zhaoyang Huang, Hujun Bao, Zhaopeng Cui, Guofeng Zhang
Abstract: Feature tracking is crucial for, structure from motion (SFM), simultaneous localization and mapping (SLAM), object tracking and various computer vision tasks. Event cameras, known for their high temporal resolution and ability to capture asynchronous changes, have gained significant attention for their potential in feature tracking, especially in challenging conditions. However, event cameras lack the fine-grained texture information that conventional cameras provide, leading to error accumulation in tracking. To address this, we propose a novel framework, BlinkTrack, which integrates event data with RGB images for high-frequency feature tracking. Our method extends the traditional Kalman filter into a learning-based framework, utilizing differentiable Kalman filters in both event and image branches. This approach improves single-modality tracking, resolves ambiguities, and supports asynchronous data fusion. We also introduce new synthetic and augmented datasets to better evaluate our model. Experimental results indicate that BlinkTrack significantly outperforms existing event-based methods, exceeding 100 FPS with preprocessed event data and 80 FPS with multi-modality data.
Authors: Ruizhe Zheng, Lichao Sun
Abstract: Decoding visual-semantic information from brain signals, such as functional MRI (fMRI), across different subjects poses significant challenges, including low signal-to-noise ratio, limited data availability, and cross-subject variability. Recent advancements in large language models (LLMs) show remarkable effectiveness in processing multimodal information. In this study, we introduce an LLM-based approach for reconstructing visual-semantic information from fMRI signals elicited by video stimuli. Specifically, we employ fine-tuning techniques on an fMRI encoder equipped with adaptors to transform brain responses into latent representations aligned with the video stimuli. Subsequently, these representations are mapped to textual modality by LLM. In particular, we integrate self-supervised domain adaptation methods to enhance the alignment between visual-semantic information and brain responses. Our proposed method achieves good results using various quantitative semantic metrics, while yielding similarity with ground-truth information.
Authors: Weng Fei Low, Gim Hee Lee
Abstract: The stark contrast in the design philosophy of an event camera makes it particularly ideal for operating under high-speed, high dynamic range and low-light conditions, where standard cameras underperform. Nonetheless, event cameras still suffer from some amount of motion blur, especially under these challenging conditions, in contrary to what most think. This is attributed to the limited bandwidth of the event sensor pixel, which is mostly proportional to the light intensity. Thus, to ensure that event cameras can truly excel in such conditions where it has an edge over standard cameras, it is crucial to account for event motion blur in downstream applications, especially reconstruction. However, none of the recent works on reconstructing Neural Radiance Fields (NeRFs) from events, nor event simulators, have considered the full effects of event motion blur. To this end, we propose, Deblur e-NeRF, a novel method to directly and effectively reconstruct blur-minimal NeRFs from motion-blurred events generated under high-speed motion or low-light conditions. The core component of this work is a physically-accurate pixel bandwidth model proposed to account for event motion blur under arbitrary speed and lighting conditions. We also introduce a novel threshold-normalized total variation loss to improve the regularization of large textureless patches. Experiments on real and novel realistically simulated sequences verify our effectiveness. Our code, event simulator and synthetic event dataset will be open-sourced.
Authors: Junchen Yu, Si-Yuan Cao, Runmin Zhang, Chenghao Zhang, Jianxin Hu, Zhu Yu, Hui-liang Shen
Abstract: We propose a novel unsupervised cross-modal homography estimation framework, based on interleaved modality transfer and self-supervised homography prediction, named InterNet. InterNet integrates modality transfer and self-supervised homography estimation, introducing an innovative interleaved optimization framework to alternately promote both components. The modality transfer gradually narrows the modality gaps, facilitating the self-supervised homography estimation to fully leverage the synthetic intra-modal data. The self-supervised homography estimation progressively achieves reliable predictions, thereby providing robust cross-modal supervision for the modality transfer. To further boost the estimation accuracy, we also formulate a fine-grained homography feature loss to improve the connection between two components. Furthermore, we employ a simple yet effective distillation training technique to reduce model parameters and improve cross-domain generalization ability while maintaining comparable performance. Experiments reveal that InterNet achieves the state-of-the-art (SOTA) performance among unsupervised methods, and even outperforms many supervised methods such as MHN and LocalTrans.
Authors: Jacopo Dapueto, Nicoletta Noceti, Francesca Odone
Abstract: Developing meaningful and efficient representations that separate the fundamental structure of the data generation mechanism is crucial in representation learning. However, Disentangled Representation Learning has not fully shown its potential on real images, because of correlated generative factors, their resolution and limited access to ground truth labels. Specifically on the latter, we investigate the possibility of leveraging synthetic data to learn general-purpose disentangled representations applicable to real data, discussing the effect of fine-tuning and what properties of disentanglement are preserved after the transfer. We provide an extensive empirical study to address these issues. In addition, we propose a new interpretable intervention-based metric, to measure the quality of factors encoding in the representation. Our results indicate that some level of disentanglement, transferring a representation from synthetic to real data, is possible and effective.
Authors: Song Wang, Zhongdao Wang, Jiawei Yu, Wentong Li, Bailan Feng, Junbo Chen, Jianke Zhu
Abstract: Vision-centric semantic occupancy prediction plays a crucial role in autonomous driving, which requires accurate and reliable predictions from low-cost sensors. Although having notably narrowed the accuracy gap with LiDAR, there is still few research effort to explore the reliability in predicting semantic occupancy from camera. In this paper, we conduct a comprehensive evaluation of existing semantic occupancy prediction models from a reliability perspective for the first time. Despite the gradual alignment of camera-based models with LiDAR in term of accuracy, a significant reliability gap persists. To addresses this concern, we propose ReliOcc, a method designed to enhance the reliability of camera-based occupancy networks. ReliOcc provides a plug-and-play scheme for existing models, which integrates hybrid uncertainty from individual voxels with sampling-based noise and relative voxels through mix-up learning. Besides, an uncertainty-aware calibration strategy is devised to further enhance model reliability in offline mode. Extensive experiments under various settings demonstrate that ReliOcc significantly enhances model reliability while maintaining the accuracy of both geometric and semantic predictions. Importantly, our proposed approach exhibits robustness to sensor failures and out of domain noises during inference.
Authors: Kai Chen, Yunhao Gou, Runhui Huang, Zhili Liu, Daxin Tan, Jing Xu, Chunwei Wang, Yi Zhu, Yihan Zeng, Kuo Yang, Dingdong Wang, Kun Xiang, Haoyuan Li, Haoli Bai, Jianhua Han, Xiaohui Li, Weike Jin, Nian Xie, Yu Zhang, James T. Kwok, Hengshuang Zhao, Xiaodan Liang, Dit-Yan Yeung, Xiao Chen, Zhenguo Li, Wei Zhang, Qun Liu, Lanqing Hong, Lu Hou, Hang Xu
Abstract: GPT-4o, an omni-modal model that enables vocal conversations with diverse emotions and tones, marks a milestone for omni-modal foundation models. However, empowering Large Language Models to perceive and generate images, texts, and speeches end-to-end with publicly available data remains challenging in the open-source community. Existing vision-language models rely on external tools for the speech processing, while speech-language models still suffer from limited or even without vision-understanding abilities. To address this gap, we propose EMOVA (EMotionally Omni-present Voice Assistant), to enable Large Language Models with end-to-end speech capabilities while maintaining the leading vision-language performance. With a semantic-acoustic disentangled speech tokenizer, we notice surprisingly that omni-modal alignment can further enhance vision-language and speech abilities compared with the corresponding bi-modal aligned counterparts. Moreover, a lightweight style module is proposed for flexible speech style controls (e.g., emotions and pitches). For the first time, EMOVA achieves state-of-the-art performance on both the vision-language and speech benchmarks, and meanwhile, supporting omni-modal spoken dialogue with vivid emotions.
Authors: Soeun Lee, Si-Woo Kim, Taewhan Kim, Dong-Jin Kim
Abstract: Recent advancements in image captioning have explored text-only training methods to overcome the limitations of paired image-text data. However, existing text-only training methods often overlook the modality gap between using text data during training and employing images during inference. To address this issue, we propose a novel approach called Image-like Retrieval, which aligns text features with visually relevant features to mitigate the modality gap. Our method further enhances the accuracy of generated captions by designing a Fusion Module that integrates retrieved captions with input features. Additionally, we introduce a Frequency-based Entity Filtering technique that significantly improves caption quality. We integrate these methods into a unified framework, which we refer to as IFCap ($\textbf{I}$mage-like Retrieval and $\textbf{F}$requency-based Entity Filtering for Zero-shot $\textbf{Cap}$tioning). Through extensive experimentation, our straightforward yet powerful approach has demonstrated its efficacy, outperforming the state-of-the-art methods by a significant margin in both image captioning and video captioning compared to zero-shot captioning based on text-only training.
Authors: Kartik Garg, Sai Shubodh Puligilla, Shishir Kolathaya, Madhava Krishna, Sourav Garg
Abstract: Accurately recognizing a revisited place is crucial for embodied agents to localize and navigate. This requires visual representations to be distinct, despite strong variations in camera viewpoint and scene appearance. Existing visual place recognition pipelines encode the "whole" image and search for matches. This poses a fundamental challenge in matching two images of the same place captured from different camera viewpoints: "the similarity of what overlaps can be dominated by the dissimilarity of what does not overlap". We address this by encoding and searching for "image segments" instead of the whole images. We propose to use open-set image segmentation to decompose an image into `meaningful' entities (i.e., things and stuff). This enables us to create a novel image representation as a collection of multiple overlapping subgraphs connecting a segment with its neighboring segments, dubbed SuperSegment. Furthermore, to efficiently encode these SuperSegments into compact vector representations, we propose a novel factorized representation of feature aggregation. We show that retrieving these partial representations leads to significantly higher recognition recall than the typical whole image based retrieval. Our segments-based approach, dubbed SegVLAD, sets a new state-of-the-art in place recognition on a diverse selection of benchmark datasets, while being applicable to both generic and task-specialized image encoders. Finally, we demonstrate the potential of our method to ``revisit anything'' by evaluating our method on an object instance retrieval task, which bridges the two disparate areas of research: visual place recognition and object-goal navigation, through their common aim of recognizing goal objects specific to a place. Source code: https://github.com/AnyLoc/Revisit-Anything.
Authors: Rwiddhi Chakraborty, Yinong Wang, Jialu Gao, Runkai Zheng, Cheng Zhang, Fernando De la Torre
Abstract: The widespread success of deep learning models today is owed to the curation of extensive datasets significant in size and complexity. However, such models frequently pick up inherent biases in the data during the training process, leading to unreliable predictions. Diagnosing and debiasing datasets is thus a necessity to ensure reliable model performance. In this paper, we present CONBIAS, a novel framework for diagnosing and mitigating Concept co-occurrence Biases in visual datasets. CONBIAS represents visual datasets as knowledge graphs of concepts, enabling meticulous analysis of spurious concept co-occurrences to uncover concept imbalances across the whole dataset. Moreover, we show that by employing a novel clique-based concept balancing strategy, we can mitigate these imbalances, leading to enhanced performance on downstream tasks. Extensive experiments show that data augmentation based on a balanced concept distribution augmented by CONBIAS improves generalization performance across multiple datasets compared to state-of-the-art methods. We will make our code and data publicly available.
Authors: Huan Wang, Feitong Tan, Ziqian Bai, Yinda Zhang, Shichen Liu, Qiangeng Xu, Menglei Chai, Anish Prabhu, Rohit Pandey, Sean Fanello, Zeng Huang, Yun Fu
Abstract: Recent works have shown that neural radiance fields (NeRFs) on top of parametric models have reached SOTA quality to build photorealistic head avatars from a monocular video. However, one major limitation of the NeRF-based avatars is the slow rendering speed due to the dense point sampling of NeRF, preventing them from broader utility on resource-constrained devices. We introduce LightAvatar, the first head avatar model based on neural light fields (NeLFs). LightAvatar renders an image from 3DMM parameters and a camera pose via a single network forward pass, without using mesh or volume rendering. The proposed approach, while being conceptually appealing, poses a significant challenge towards real-time efficiency and training stability. To resolve them, we introduce dedicated network designs to obtain proper representations for the NeLF model and maintain a low FLOPs budget. Meanwhile, we tap into a distillation-based training strategy that uses a pretrained avatar model as teacher to synthesize abundant pseudo data for training. A warping field network is introduced to correct the fitting error in the real data so that the model can learn better. Extensive experiments suggest that our method can achieve new SOTA image quality quantitatively or qualitatively, while being significantly faster than the counterparts, reporting 174.1 FPS (512x512 resolution) on a consumer-grade GPU (RTX3090) with no customized optimization.
Authors: Runze He, Kai Ma, Linjiang Huang, Shaofei Huang, Jialin Gao, Xiaoming Wei, Jiao Dai, Jizhong Han, Si Liu
Abstract: Introducing user-specified visual concepts in image editing is highly practical as these concepts convey the user's intent more precisely than text-based descriptions. We propose FreeEdit, a novel approach for achieving such reference-based image editing, which can accurately reproduce the visual concept from the reference image based on user-friendly language instructions. Our approach leverages the multi-modal instruction encoder to encode language instructions to guide the editing process. This implicit way of locating the editing area eliminates the need for manual editing masks. To enhance the reconstruction of reference details, we introduce the Decoupled Residual ReferAttention (DRRA) module. This module is designed to integrate fine-grained reference features extracted by a detail extractor into the image editing process in a residual way without interfering with the original self-attention. Given that existing datasets are unsuitable for reference-based image editing tasks, particularly due to the difficulty in constructing image triplets that include a reference image, we curate a high-quality dataset, FreeBench, using a newly developed twice-repainting scheme. FreeBench comprises the images before and after editing, detailed editing instructions, as well as a reference image that maintains the identity of the edited object, encompassing tasks such as object addition, replacement, and deletion. By conducting phased training on FreeBench followed by quality tuning, FreeEdit achieves high-quality zero-shot editing through convenient language instructions. We conduct extensive experiments to evaluate the effectiveness of FreeEdit across multiple task types, demonstrating its superiority over existing methods. The code will be available at: https://freeedit.github.io/.
Authors: Mirela Ostrek, Justus Thies
Abstract: Rapid advances in the field of generative AI and text-to-image methods in particular have transformed the way we interact with and perceive computer-generated imagery today. In parallel, much progress has been made in 3D face reconstruction, using 3D Morphable Models (3DMM). In this paper, we present SVP, a novel hybrid 2D/3D generation method that outputs photorealistic videos of talking faces leveraging a large pre-trained text-to-image prior (2D), controlled via a 3DMM (3D). Specifically, we introduce a person-specific fine-tuning of a general 2D stable diffusion model which we lift to a video model by providing temporal 3DMM sequences as conditioning and by introducing a temporal denoising procedure. As an output, this model generates temporally smooth imagery of a person with 3DMM-based controls, i.e., a person-specific avatar. The facial appearance of this person-specific avatar can be edited and morphed to text-defined celebrities, without any fine-tuning at test time. The method is analyzed quantitatively and qualitatively, and we show that our method outperforms state-of-the-art monocular head avatar methods.
Authors: Helin Cao, Sven Behnke
Abstract: Perception systems play a crucial role in autonomous driving, incorporating multiple sensors and corresponding computer vision algorithms. 3D LiDAR sensors are widely used to capture sparse point clouds of the vehicle's surroundings. However, such systems struggle to perceive occluded areas and gaps in the scene due to the sparsity of these point clouds and their lack of semantics. To address these challenges, Semantic Scene Completion (SSC) jointly predicts unobserved geometry and semantics in the scene given raw LiDAR measurements, aiming for a more complete scene representation. Building on promising results of diffusion models in image generation and super-resolution tasks, we propose their extension to SSC by implementing the noising and denoising diffusion processes in the point and semantic spaces individually. To control the generation, we employ semantic LiDAR point clouds as conditional input and design local and global regularization losses to stabilize the denoising process. We evaluate our approach on autonomous driving datasets and our approach outperforms the state-of-the-art for SSC.
Authors: Abid Hasan Zim, Aquib Iqbal, Zaid Al-Huda, Asad Malik, Minoru Kuribayash
Abstract: Crack detection, particularly from pavement images, presents a formidable challenge in the domain of computer vision due to several inherent complexities such as intensity inhomogeneity, intricate topologies, low contrast, and noisy backgrounds. Automated crack detection is crucial for maintaining the structural integrity of essential infrastructures, including buildings, pavements, and bridges. Existing lightweight methods often face challenges including computational inefficiency, complex crack patterns, and difficult backgrounds, leading to inaccurate detection and impracticality for real-world applications. To address these limitations, we propose EfficientCrackNet, a lightweight hybrid model combining Convolutional Neural Networks (CNNs) and transformers for precise crack segmentation. EfficientCrackNet integrates depthwise separable convolutions (DSC) layers and MobileViT block to capture both global and local features. The model employs an Edge Extraction Method (EEM) and for efficient crack edge detection without pretraining, and Ultra-Lightweight Subspace Attention Module (ULSAM) to enhance feature extraction. Extensive experiments on three benchmark datasets Crack500, DeepCrack, and GAPs384 demonstrate that EfficientCrackNet achieves superior performance compared to existing lightweight models, while requiring only 0.26M parameters, and 0.483 FLOPs (G). The proposed model offers an optimal balance between accuracy and computational efficiency, outperforming state-of-the-art lightweight models, and providing a robust and adaptable solution for real-world crack segmentation.
Authors: Rob A. J. de Mooij, Josien P. W. Pluim, Cian M. Scannell
Abstract: Self-supervised pretraining (SSP) has shown promising results in learning from large unlabeled datasets and, thus, could be useful for automated cardiovascular magnetic resonance (CMR) short-axis cine segmentation. However, inconsistent reports of the benefits of SSP for segmentation have made it difficult to apply SSP to CMR. Therefore, this study aimed to evaluate SSP methods for CMR cine segmentation. To this end, short-axis cine stacks of 296 subjects (90618 2D slices) were used for unlabeled pretraining with four SSP methods; SimCLR, positional contrastive learning, DINO, and masked image modeling (MIM). Subsets of varying numbers of subjects were used for supervised fine-tuning of 2D models for each SSP method, as well as to train a 2D baseline model from scratch. The fine-tuned models were compared to the baseline using the 3D Dice similarity coefficient (DSC) in a test dataset of 140 subjects. The SSP methods showed no performance gains with the largest supervised fine-tuning subset compared to the baseline (DSC = 0.89). When only 10 subjects (231 2D slices) are available for supervised training, SSP using MIM (DSC = 0.86) improves over training from scratch (DSC = 0.82). This study found that SSP is valuable for CMR cine segmentation when labeled training data is scarce, but does not aid state-of-the-art deep learning methods when ample labeled data is available. Moreover, the choice of SSP method is important. The code is publicly available at: https://github.com/q-cardIA/ssp-cmr-cine-segmentation
Authors: Alvaro Patricio, Joao Valente, Atabak Dehban, Ines Cadilha, Daniel Reis, Rodrigo Ventura
Abstract: The integration of Artificial Intelligence (AI) and Augmented Reality (AR) is set to transform satellite Assembly, Integration, and Testing (AIT) processes by enhancing precision, minimizing human error, and improving operational efficiency in cleanroom environments. This paper presents a technical description of the European Space Agency's (ESA) project "AI for AR in Satellite AIT," which combines real-time computer vision and AR systems to assist technicians during satellite assembly. Leveraging Microsoft HoloLens 2 as the AR interface, the system delivers context-aware instructions and real-time feedback, tackling the complexities of object recognition and 6D pose estimation in AIT workflows. All AI models demonstrated over 70% accuracy, with the detection model exceeding 95% accuracy, indicating a high level of performance and reliability. A key contribution of this work lies in the effective use of synthetic data for training AI models in AR applications, addressing the significant challenges of obtaining real-world datasets in highly dynamic satellite environments, as well as the creation of the Segmented Anything Model for Automatic Labelling (SAMAL), which facilitates the automatic annotation of real data, achieving speeds up to 20 times faster than manual human annotation. The findings demonstrate the efficacy of AI-driven AR systems in automating critical satellite assembly tasks, setting a foundation for future innovations in the space industry.
Authors: Lucia Gordon, Nikhil Behari, Samuel Collier, Elizabeth Bondi-Kelly, Jackson A. Killian, Catherine Ressijac, Peter Boucher, Andrew Davies, Milind Tambe
Abstract: Much of Earth's charismatic megafauna is endangered by human activities, particularly the rhino, which is at risk of extinction due to the poaching crisis in Africa. Monitoring rhinos' movement is crucial to their protection but has unfortunately proven difficult because rhinos are elusive. Therefore, instead of tracking rhinos, we propose the novel approach of mapping communal defecation sites, called middens, which give information about rhinos' spatial behavior valuable to anti-poaching, management, and reintroduction efforts. This paper provides the first-ever mapping of rhino midden locations by building classifiers to detect them using remotely sensed thermal, RGB, and LiDAR imagery in passive and active learning settings. As existing active learning methods perform poorly due to the extreme class imbalance in our dataset, we design MultimodAL, an active learning system employing a ranking technique and multimodality to achieve competitive performance with passive learning models with 94% fewer labels. Our methods could therefore save over 76 hours in labeling time when used on a similarly-sized dataset. Unexpectedly, our midden map reveals that rhino middens are not randomly distributed throughout the landscape; rather, they are clustered. Consequently, rangers should be targeted at areas with high midden densities to strengthen anti-poaching efforts, in line with UN Target 15.7.
Authors: Ye Liu, Zongyang Ma, Zhongang Qi, Yang Wu, Ying Shan, Chang Wen Chen
Abstract: Recent advances in Video Large Language Models (Video-LLMs) have demonstrated their great potential in general-purpose video understanding. To verify the significance of these models, a number of benchmarks have been proposed to diagnose their capabilities in different scenarios. However, existing benchmarks merely evaluate models through video-level question-answering, lacking fine-grained event-level assessment and task diversity. To fill this gap, we introduce E.T. Bench (Event-Level & Time-Sensitive Video Understanding Benchmark), a large-scale and high-quality benchmark for open-ended event-level video understanding. Categorized within a 3-level task taxonomy, E.T. Bench encompasses 7.3K samples under 12 tasks with 7K videos (251.4h total length) under 8 domains, providing comprehensive evaluations. We extensively evaluated 8 Image-LLMs and 12 Video-LLMs on our benchmark, and the results reveal that state-of-the-art models for coarse-level (video-level) understanding struggle to solve our fine-grained tasks, e.g., grounding event-of-interests within videos, largely due to the short video context length, improper time representations, and lack of multi-event training data. Focusing on these issues, we further propose a strong baseline model, E.T. Chat, together with an instruction-tuning dataset E.T. Instruct 164K tailored for fine-grained event-level understanding. Our simple but effective solution demonstrates superior performance in multiple scenarios.
Authors: Jiaxiang Tang, Zhaoshuo Li, Zekun Hao, Xian Liu, Gang Zeng, Ming-Yu Liu, Qinsheng Zhang
Abstract: Current auto-regressive mesh generation methods suffer from issues such as incompleteness, insufficient detail, and poor generalization. In this paper, we propose an Auto-regressive Auto-encoder (ArAE) model capable of generating high-quality 3D meshes with up to 4,000 faces at a spatial resolution of $512^3$. We introduce a novel mesh tokenization algorithm that efficiently compresses triangular meshes into 1D token sequences, significantly enhancing training efficiency. Furthermore, our model compresses variable-length triangular meshes into a fixed-length latent space, enabling training latent diffusion models for better generalization. Extensive experiments demonstrate the superior quality, diversity, and generalization capabilities of our model in both point cloud and image-conditioned mesh generation tasks.
Authors: Yuexi Du, John Onofrey, Nicha C. Dvornek
Abstract: Contrastive Language-Image Pre-training (CLIP) shows promise in medical image analysis but requires substantial data and computational resources. Due to these restrictions, existing CLIP applications in medical imaging focus mainly on modalities like chest X-rays that have abundant image-report data available, leaving many other important modalities under-explored. Here, we propose the first adaptation of the full CLIP model to mammography, which presents significant challenges due to labeled data scarcity, high-resolution images with small regions of interest, and data imbalance. We first develop a specialized supervision framework for mammography that leverages its multi-view nature. Furthermore, we design a symmetric local alignment module to better focus on detailed features in high-resolution images. Lastly, we incorporate a parameter-efficient fine-tuning approach for large language models pre-trained with medical knowledge to address data limitations. Our multi-view and multi-scale alignment (MaMA) method outperforms state-of-the-art baselines for three different tasks on two large real-world mammography datasets, EMBED and RSNA-Mammo, with only 52% model size compared with the largest baseline.
Authors: Jing He, Haodong Li, Wei Yin, Yixun Liang, Leheng Li, Kaiqiang Zhou, Hongbo Liu, Bingbing Liu, Ying-Cong Chen
Abstract: Leveraging the visual priors of pre-trained text-to-image diffusion models offers a promising solution to enhance zero-shot generalization in dense prediction tasks. However, existing methods often uncritically use the original diffusion formulation, which may not be optimal due to the fundamental differences between dense prediction and image generation. In this paper, we provide a systemic analysis of the diffusion formulation for the dense prediction, focusing on both quality and efficiency. And we find that the original parameterization type for image generation, which learns to predict noise, is harmful for dense prediction; the multi-step noising/denoising diffusion process is also unnecessary and challenging to optimize. Based on these insights, we introduce Lotus, a diffusion-based visual foundation model with a simple yet effective adaptation protocol for dense prediction. Specifically, Lotus is trained to directly predict annotations instead of noise, thereby avoiding harmful variance. We also reformulate the diffusion process into a single-step procedure, simplifying optimization and significantly boosting inference speed. Additionally, we introduce a novel tuning strategy called detail preserver, which achieves more accurate and fine-grained predictions. Without scaling up the training data or model capacity, Lotus achieves SoTA performance in zero-shot depth and normal estimation across various datasets. It also significantly enhances efficiency, being hundreds of times faster than most existing diffusion-based methods.
Authors: Chenming Zhu, Tai Wang, Wenwei Zhang, Jiangmiao Pang, Xihui Liu
Abstract: Recent advancements in Large Multimodal Models (LMMs) have greatly enhanced their proficiency in 2D visual understanding tasks, enabling them to effectively process and understand images and videos. However, the development of LMMs with 3D-awareness for 3D scene understanding has been hindered by the lack of large-scale 3D vision-language datasets and powerful 3D encoders. In this paper, we introduce a simple yet effective framework called LLaVA-3D. Leveraging the strong 2D understanding priors from LLaVA, our LLaVA-3D efficiently adapts LLaVA for 3D scene understanding without compromising 2D understanding capabilities. To achieve this, we employ a simple yet effective representation, 3D Patch, which connects 2D CLIP patch features with their corresponding positions in 3D space. By integrating the 3D Patches into 2D LMMs and employing joint 2D and 3D vision-language instruction tuning, we establish a unified architecture for both 2D image understanding and 3D scene understanding. Experimental results show that LLaVA-3D converges 3.5x faster than existing 3D LMMs when trained on 3D vision-language datasets. Moreover, LLaVA-3D not only achieves state-of-the-art performance across various 3D tasks but also maintains comparable 2D image understanding and vision-language conversation capabilities with LLaVA.
Authors: Fangzhou Hong, Vladimir Guzov, Hyo Jin Kim, Yuting Ye, Richard Newcombe, Ziwei Liu, Lingni Ma
Abstract: As the prevalence of wearable devices, learning egocentric motions becomes essential to develop contextual AI. In this work, we present EgoLM, a versatile framework that tracks and understands egocentric motions from multi-modal inputs, e.g., egocentric videos and motion sensors. EgoLM exploits rich contexts for the disambiguation of egomotion tracking and understanding, which are ill-posed under single modality conditions. To facilitate the versatile and multi-modal framework, our key insight is to model the joint distribution of egocentric motions and natural languages using large language models (LLM). Multi-modal sensor inputs are encoded and projected to the joint latent space of language models, and used to prompt motion generation or text generation for egomotion tracking or understanding, respectively. Extensive experiments on large-scale multi-modal human motion dataset validate the effectiveness of EgoLM as a generalist model for universal egocentric learning.
Authors: Wenliang Zhao, Minglei Shi, Xumin Yu, Jie Zhou, Jiwen Lu
Abstract: Building on the success of diffusion models in visual generation, flow-based models reemerge as another prominent family of generative models that have achieved competitive or better performance in terms of both visual quality and inference speed. By learning the velocity field through flow-matching, flow-based models tend to produce a straighter sampling trajectory, which is advantageous during the sampling process. However, unlike diffusion models for which fast samplers are well-developed, efficient sampling of flow-based generative models has been rarely explored. In this paper, we propose a framework called FlowTurbo to accelerate the sampling of flow-based models while still enhancing the sampling quality. Our primary observation is that the velocity predictor's outputs in the flow-based models will become stable during the sampling, enabling the estimation of velocity via a lightweight velocity refiner. Additionally, we introduce several techniques including a pseudo corrector and sample-aware compilation to further reduce inference time. Since FlowTurbo does not change the multi-step sampling paradigm, it can be effectively applied for various tasks such as image editing, inpainting, etc. By integrating FlowTurbo into different flow-based models, we obtain an acceleration ratio of 53.1%$\sim$58.3% on class-conditional generation and 29.8%$\sim$38.5% on text-to-image generation. Notably, FlowTurbo reaches an FID of 2.12 on ImageNet with 100 (ms / img) and FID of 3.93 with 38 (ms / img), achieving the real-time image generation and establishing the new state-of-the-art. Code is available at https://github.com/shiml20/FlowTurbo.
Authors: Th\'eo Bodrito, Olivier Flasseur, Julien Mairal, Jean Ponce, Maud Langlois, Anne-Marie Lagrange
Abstract: Direct imaging of exoplanets is particularly challenging due to the high contrast between the planet and the star luminosities, and their small angular separation. In addition to tailored instrumental facilities implementing adaptive optics and coronagraphy, post-processing methods combining several images recorded in pupil tracking mode are needed to attenuate the nuisances corrupting the signals of interest. Most of these post-processing methods build a model of the nuisances from the target observations themselves, resulting in strongly limited detection sensitivity at short angular separations due to the lack of angular diversity. To address this issue, we propose to build the nuisance model from an archive of multiple observations by leveraging supervised deep learning techniques. The proposed approach casts the detection problem as a reconstruction task and captures the structure of the nuisance from two complementary representations of the data. Unlike methods inspired by reference differential imaging, the proposed model is highly non-linear and does not resort to explicit image-to-image similarity measurements and subtractions. The proposed approach also encompasses statistical modeling of learnable spatial features. The latter is beneficial to improve both the detection sensitivity and the robustness against heterogeneous data. We apply the proposed algorithm to several datasets from the VLT/SPHERE instrument, and demonstrate a superior precision-recall trade-off compared to the PACO algorithm. Interestingly, the gain is especially important when the diversity induced by ADI is the most limited, thus supporting the ability of the proposed approach to learn information across multiple observations.
Authors: Marcos V Conde, Zhijun Lei, Wen Li, Christos Bampis, Ioannis Katsavounidis, Radu Timofte
Abstract: Video super-resolution (VSR) is a critical task for enhancing low-bitrate and low-resolution videos, particularly in streaming applications. While numerous solutions have been developed, they often suffer from high computational demands, resulting in low frame rates (FPS) and poor power efficiency, especially on mobile platforms. In this work, we compile different methods to address these challenges, the solutions are end-to-end real-time video super-resolution frameworks optimized for both high performance and low runtime. We also introduce a new test set of high-quality 4K videos to further validate the approaches. The proposed solutions tackle video up-scaling for two applications: 540p to 4K (x4) as a general case, and 360p to 1080p (x3) more tailored towards mobile devices. In both tracks, the solutions have a reduced number of parameters and operations (MACs), allow high FPS, and improve VMAF and PSNR over interpolation baselines. This report gauges some of the most efficient video super-resolution methods to date.
Authors: Pandiyaraju V, Shravan Venkatraman, Abeshek A, Aravintakshan S A, Pavan Kumar S, Madhan S
Abstract: Tumors in the brain result from abnormal cell growth within the brain tissue, arising from various types of brain cells. When left undiagnosed, they lead to severe neurological deficits such as cognitive impairment, motor dysfunction, and sensory loss. As the tumor grows, it causes an increase in intracranial pressure, potentially leading to life-threatening complications such as brain herniation. Therefore, early detection and treatment are necessary to manage the complications caused by such tumors to slow down their growth. Numerous works involving deep learning (DL) and artificial intelligence (AI) are being carried out to assist physicians in early diagnosis by utilizing the scans obtained through Magnetic Resonance Imaging (MRI). Our research proposes DL frameworks for localizing, segmenting, and classifying the grade of these gliomas from MRI images to solve this critical issue. In our localization framework, we enhance the LinkNet framework with a VGG19- inspired encoder architecture for improved multimodal tumor feature extraction, along with spatial and graph attention mechanisms to refine feature focus and inter-feature relationships. Following this, we integrated the SeResNet101 CNN model as the encoder backbone into the LinkNet framework for tumor segmentation, which achieved an IoU Score of 96%. To classify the segmented tumors, we combined the SeResNet152 feature extractor with an Adaptive Boosting classifier, which yielded an accuracy of 98.53%. Our proposed models demonstrated promising results, with the potential to advance medical AI by enabling early diagnosis and providing more accurate treatment options for patients.
Authors: Jingwen Lyu, Yue Wang, Christian Pedersen, Spencer Jones, Dhruv Balwada
Abstract: Knowledge of ocean circulation is important for understanding and predicting weather and climate, and managing the blue economy. This circulation can be estimated through Sea Surface Height (SSH) observations, but requires decomposing the SSH into contributions from balanced and unbalanced motions (BMs and UBMs). This decomposition is particularly pertinent for the novel SWOT satellite, which measures SSH at an unprecedented spatial resolution. Specifically, the requirement, and the goal of this work, is to decompose instantaneous SSH into BMs and UBMs. While a few studies using deep learning (DL) approaches have shown promise in framing this decomposition as an image-to-image translation task, these models struggle to work well across a wide range of spatial scales and require extensive training data, which is scarce in this domain. These challenges are not unique to our task, and pervade many problems requiring multi-scale fidelity. We show that these challenges can be addressed by using zero-phase component analysis (ZCA) whitening and data augmentation; making this a viable option for SSH decomposition across scales.
Authors: Alif Bin Abdul Qayyum, Xihaier Luo, Nathan M. Urban, Xiaoning Qian, Byung-Jun Yoon
Abstract: The world is moving towards clean and renewable energy sources, such as wind energy, in an attempt to reduce greenhouse gas emissions that contribute to global warming. To enhance the analysis and storage of wind data, we introduce a deep learning framework designed to simultaneously enable effective dimensionality reduction and continuous representation of multi-altitude wind data from discrete observations. The framework consists of three key components: dimensionality reduction, cross-modal prediction, and super-resolution. We aim to: (1) improve data resolution across diverse climatic conditions to recover high-resolution details; (2) reduce data dimensionality for more efficient storage of large climate datasets; and (3) enable cross-prediction between wind data measured at different heights. Comprehensive testing confirms that our approach surpasses existing methods in both super-resolution quality and compression efficiency.
Authors: Ce Zhou (Michigan State University), Qiben Yan (Michigan State University), Daniel Kent (Michigan State University), Guangjing Wang (Michigan State University), Ziqi Zhang (Peking University), Hayder Radha (Michigan State University)
Abstract: Monocular Depth Estimation (MDE) plays a crucial role in vision-based Autonomous Driving (AD) systems. It utilizes a single-camera image to determine the depth of objects, facilitating driving decisions such as braking a few meters in front of a detected obstacle or changing lanes to avoid collision. In this paper, we investigate the security risks associated with monocular vision-based depth estimation algorithms utilized by AD systems. By exploiting the vulnerabilities of MDE and the principles of optical lenses, we introduce LensAttack, a physical attack that involves strategically placing optical lenses on the camera of an autonomous vehicle to manipulate the perceived object depths. LensAttack encompasses two attack formats: concave lens attack and convex lens attack, each utilizing different optical lenses to induce false depth perception. We begin by constructing a mathematical model of our attack, incorporating various attack parameters. Subsequently, we simulate the attack and evaluate its real-world performance in driving scenarios to demonstrate its effect on state-of-the-art MDE models. The results highlight the significant impact of LensAttack on the accuracy of depth estimation in AD systems.
Authors: Ruining Yang, Lili Su
Abstract: Modern vehicles are equipped with multiple information-collection devices such as sensors and cameras, continuously generating a large volume of raw data. Accurately predicting the trajectories of neighboring vehicles is a vital component in understanding the complex driving environment. Yet, training trajectory prediction models is challenging in two ways. Processing the large-scale data is computation-intensive. Moreover, easy-medium driving scenarios often overwhelmingly dominate the dataset, leaving challenging driving scenarios such as dense traffic under-represented. For example, in the Argoverse motion prediction dataset, there are very few instances with $\ge 50$ agents, while scenarios with $10 \thicksim 20$ agents are far more common. In this paper, to mitigate data redundancy in the over-represented driving scenarios and to reduce the bias rooted in the data scarcity of complex ones, we propose a novel data-efficient training method based on coreset selection. This method strategically selects a small but representative subset of data while balancing the proportions of different scenario difficulties. To the best of our knowledge, we are the first to introduce a method capable of effectively condensing large-scale trajectory dataset, while achieving a state-of-the-art compression ratio. Notably, even when using only 50% of the Argoverse dataset, the model can be trained with little to no decline in performance. Moreover, the selected coreset maintains excellent generalization ability.
Authors: Ce Zhou, Qiben Yan, Sijia Liu
Abstract: Object detection is a crucial task in autonomous driving. While existing research has proposed various attacks on object detection, such as those using adversarial patches or stickers, the exploration of projection attacks on 3D surfaces remains largely unexplored. Compared to adversarial patches or stickers, which have fixed adversarial patterns, projection attacks allow for transient modifications to these patterns, enabling a more flexible attack. In this paper, we introduce an adversarial 3D projection attack specifically targeting object detection in autonomous driving scenarios. We frame the attack formulation as an optimization problem, utilizing a combination of color mapping and geometric transformation models. Our results demonstrate the effectiveness of the proposed attack in deceiving YOLOv3 and Mask R-CNN in physical settings. Evaluations conducted in an indoor environment show an attack success rate of up to 100% under low ambient light conditions, highlighting the potential damage of our attack in real-world driving scenarios.
Authors: Yongrok Kim, Junha Shin, Juhyun Lee, Hyunsuk Ko
Abstract: To display low-quality broadcast content on high-resolution screens in full-screen format, the application of Super-Resolution (SR), a key consumer technology, is essential. Recently, SR methods have been developed that not only increase resolution while preserving the original image information but also enhance the perceived quality. However, evaluating the quality of SR images generated from low-quality sources, such as SR-enhanced broadcast content, is challenging due to the need to consider both distortions and improvements. Additionally, assessing SR image quality without original high-quality sources presents another significant challenge. Unfortunately, there has been a dearth of research specifically addressing the Image Quality Assessment (IQA) of SR images under these conditions. In this work, we introduce a new IQA dataset for SR broadcast images in both 2K and 4K resolutions. We conducted a subjective quality evaluation to obtain the Mean Opinion Score (MOS) for these SR images and performed a comprehensive human study to identify the key factors influencing the perceived quality. Finally, we evaluated the performance of existing IQA metrics on our dataset. This study reveals the limitations of current metrics, highlighting the need for a more robust IQA metric that better correlates with the perceived quality of SR images.
Authors: Meng Wang (School of Electronic,Information Engineering Liaoning Technical University Xingcheng City, Liaoning Province, P. R. China), Yarong Feng (School of Electronic,Information Engineering Liaoning Technical University Xingcheng City, Liaoning Province, P. R. China), Yongwei Tang (School of Electronic,Information Engineering Liaoning Technical University Xingcheng City, Liaoning Province, P. R. China), Tian Zhang (Software college Northeastern University Shenyang, Liaoning Province, P. R. China), Yuxin Liang (School of Electronic,Information Engineering Liaoning Technical University Xingcheng City, Liaoning Province, P. R. China), Chao Lv (Department of General Surgery, Shengjing Hospital China Medical University Shenyang, Liaoning Province, P. R. China)
Abstract: Emerging of visual language models, such as the segment anything model (SAM), have made great breakthroughs in the field of universal semantic segmentation and significantly aid the improvements of medical image segmentation, in particular with the help of Medical SAM adaptor (Med-SA). However, Med-SA still can be improved, as it fine-tunes SAM in a partial adaption manner. To resolve this problem, we present a novel global medical SAM adaptor (GMed-SA) with full adaption, which can adapt SAM globally. We further combine GMed-SA and Med-SA to propose a global-local medical SAM adaptor (GLMed-SA) to adapt SAM both globally and locally. Extensive experiments have been performed on the challenging public 2D melanoma segmentation dataset. The results show that GLMed-SA outperforms several state-of-the-art semantic segmentation methods on various evaluation metrics, demonstrating the superiority of our methods.
Authors: Wenhui Dong, Bo Du, Yongchao Xu
Abstract: Many medical image segmentation methods have achieved impressive results. Yet, most existing methods do not take into account the shape-intensity prior information. This may lead to implausible segmentation results, in particular for images of unseen datasets. In this paper, we propose a novel approach to incorporate joint shape-intensity prior information into the segmentation network. Specifically, we first train a segmentation network (regarded as the teacher network) on class-wise averaged training images to extract valuable shape-intensity information, which is then transferred to a student segmentation network with the same network architecture as the teacher via knowledge distillation. In this way, the student network regarded as the final segmentation model can effectively integrate the shape-intensity prior information, yielding more accurate segmentation results. Despite its simplicity, experiments on five medical image segmentation tasks of different modalities demonstrate that the proposed Shape-Intensity Knowledge Distillation (SIKD) consistently improves several baseline models (including recent MaxStyle and SAMed) under intra-dataset evaluation, and significantly improves the cross-dataset generalization ability. The code is available at https://github.com/whdong-whu/SIKD.
Authors: Ziquan Wei, Tingting Dan, Jiaqi Ding, Paul J Laurienti, Guorong Wu
Abstract: Although modern imaging technologies allow us to study connectivity between two distinct brain regions in-vivo, an in-depth understanding of how anatomical structure supports brain function and how spontaneous functional fluctuations emerge remarkable cognition is still elusive. Meanwhile, tremendous efforts have been made in the realm of machine learning to establish the nonlinear mapping between neuroimaging data and phenotypic traits. However, the absence of neuroscience insight in the current approaches poses significant challenges in understanding cognitive behavior from transient neural activities. To address this challenge, we put the spotlight on the coupling mechanism of structural connectivity (SC) and functional connectivity (FC) by formulating such network neuroscience question into an expressive graph representation learning problem for high-order topology. Specifically, we introduce the concept of topological detour to characterize how a ubiquitous instance of FC (direct link) is supported by neural pathways (detour) physically wired by SC, which forms a cyclic loop interacted by brain structure and function. In the clich\'e of machine learning, the multi-hop detour pathway underlying SC-FC coupling allows us to devise a novel multi-head self-attention mechanism within Transformer to capture multi-modal feature representation from paired graphs of SC and FC. Taken together, we propose a biological-inspired deep model, coined as NeuroPath, to find putative connectomic feature representations from the unprecedented amount of neuroimages, which can be plugged into various downstream applications such as task recognition and disease diagnosis. We have evaluated NeuroPath on large-scale public datasets including HCP and UK Biobank under supervised and zero-shot learning, where the state-of-the-art performance by our NeuroPath indicates great potential in network neuroscience.
Authors: Kento Kawaharazuka, Yoshiki Obinata, Naoaki Kanazawa, Kei Okada, Masayuki Inaba
Abstract: In order for robots to autonomously navigate and operate in diverse environments, it is essential for them to recognize the state of their environment. On the other hand, the environmental state recognition has traditionally involved distinct methods tailored to each state to be recognized. In this study, we perform a unified environmental state recognition for robots through the spoken language with pre-trained large-scale vision-language models. We apply Visual Question Answering and Image-to-Text Retrieval, which are tasks of Vision-Language Models. We show that with our method, it is possible to recognize not only whether a room door is open/closed, but also whether a transparent door is open/closed and whether water is running in a sink, without training neural networks or manual programming. In addition, the recognition accuracy can be improved by selecting appropriate texts from the set of prepared texts based on black-box optimization. For each state recognition, only the text set and its weighting need to be changed, eliminating the need to prepare multiple different models and programs, and facilitating the management of source code and computer resource. We experimentally demonstrate the effectiveness of our method and apply it to the recognition behavior on a mobile robot, Fetch.
Authors: Kunyu Peng, Di Wen, Kailun Yang, Ao Luo, Yufan Chen, Jia Fu, M. Saquib Sarfraz, Alina Roitberg, Rainer Stiefelhagen
Abstract: In Open-Set Domain Generalization (OSDG), the model is exposed to both new variations of data appearance (domains) and open-set conditions, where both known and novel categories are present at test time. The challenges of this task arise from the dual need to generalize across diverse domains and accurately quantify category novelty, which is critical for applications in dynamic environments. Recently, meta-learning techniques have demonstrated superior results in OSDG, effectively orchestrating the meta-train and -test tasks by employing varied random categories and predefined domain partition strategies. These approaches prioritize a well-designed training schedule over traditional methods that focus primarily on data augmentation and the enhancement of discriminative feature learning. The prevailing meta-learning models in OSDG typically utilize a predefined sequential domain scheduler to structure data partitions. However, a crucial aspect that remains inadequately explored is the influence brought by strategies of domain schedulers during training. In this paper, we observe that an adaptive domain scheduler benefits more in OSDG compared with prefixed sequential and random domain schedulers. We propose the Evidential Bi-Level Hardest Domain Scheduler (EBiL-HaDS) to achieve an adaptive domain scheduler. This method strategically sequences domains by assessing their reliabilities in utilizing a follower network, trained with confidence scores learned in an evidential manner, regularized by max rebiasing discrepancy, and optimized in a bi-level manner. The results show that our method substantially improves OSDG performance and achieves more discriminative embeddings for both the seen and unseen categories. The source code will be available at https://github.com/KPeng9510/EBiL-HaDS.
Authors: Peiyong Wang, Casey. R. Myers, Lloyd C. L. Hollenberg, Udaya Parampalli
Abstract: Artificial Intelligence (AI), with its multiplier effect and wide applications in multiple areas, could potentially be an important application of quantum computing. Since modern AI systems are often built on neural networks, the design of quantum neural networks becomes a key challenge in integrating quantum computing into AI. To provide a more fine-grained characterisation of the impact of quantum components on the performance of neural networks, we propose a framework where classical neural network layers are gradually replaced by quantum layers that have the same type of input and output while keeping the flow of information between layers unchanged, different from most current research in quantum neural network, which favours an end-to-end quantum model. We start with a simple three-layer classical neural network without any normalisation layers or activation functions, and gradually change the classical layers to the corresponding quantum versions. We conduct numerical experiments on image classification datasets such as the MNIST, FashionMNIST and CIFAR-10 datasets to demonstrate the change of performance brought by the systematic introduction of quantum components. Through this framework, our research sheds new light on the design of future quantum neural network models where it could be more favourable to search for methods and frameworks that harness the advantages from both the classical and quantum worlds.
Authors: Zhangpu Li, Changhong Zou, Suxue Ma, Zhicheng Yang, Chen Du, Youbao Tang, Zhenjie Cao, Ning Zhang, Jui-Hsin Lai, Ruei-Sung Lin, Yuan Ni, Xingzhi Sun, Jing Xiao, Kai Zhang, Mei Han
Abstract: The rocketing prosperity of large language models (LLMs) in recent years has boosted the prevalence of vision-language models (VLMs) in the medical sector. In our online medical consultation scenario, a doctor responds to the texts and images provided by a patient in multiple rounds to diagnose her/his health condition, forming a multi-turn multimodal medical dialogue format. Unlike high-quality images captured by professional equipment in traditional medical visual question answering (Med-VQA), the images in our case are taken by patients' mobile phones. These images have poor quality control, with issues such as excessive background elements and the lesion area being significantly off-center, leading to degradation of vision-language alignment in the model training phase. In this paper, we propose ZALM3, a Zero-shot strategy to improve vision-language ALignment in Multi-turn Multimodal Medical dialogue. Since we observe that the preceding text conversations before an image can infer the regions of interest (RoIs) in the image, ZALM3 employs an LLM to summarize the keywords from the preceding context and a visual grounding model to extract the RoIs. The updated images eliminate unnecessary background noise and provide more effective vision-language alignment. To better evaluate our proposed method, we design a new subjective assessment metric for multi-turn unimodal/multimodal medical dialogue to provide a fine-grained performance comparison. Our experiments across three different clinical departments remarkably demonstrate the efficacy of ZALM3 with statistical significance.
Authors: Jiawei Du, Xin Zhang, Juncheng Hu, Wenxin Huang, Joey Tianyi Zhou
Abstract: The sharp increase in data-related expenses has motivated research into condensing datasets while retaining the most informative features. Dataset distillation has thus recently come to the fore. This paradigm generates synthetic dataset that are representative enough to replace the original dataset in training a neural network. To avoid redundancy in these synthetic datasets, it is crucial that each element contains unique features and remains diverse from others during the synthesis stage. In this paper, we provide a thorough theoretical and empirical analysis of diversity within synthesized datasets. We argue that enhancing diversity can improve the parallelizable yet isolated synthesizing approach. Specifically, we introduce a novel method that employs dynamic and directed weight adjustment techniques to modulate the synthesis process, thereby maximizing the representativeness and diversity of each synthetic instance. Our method ensures that each batch of synthetic data mirrors the characteristics of a large, varying subset of the original dataset. Extensive experiments across multiple datasets, including CIFAR, Tiny-ImageNet, and ImageNet-1K, demonstrate the superior performance of our method, highlighting its effectiveness in producing diverse and representative synthetic datasets with minimal computational expense.
Authors: Joakim Tutt, Slava Voloshynovskiy
Abstract: Copy Detection Patterns (CDPs) are crucial elements in modern security applications, playing a vital role in safeguarding industries such as food, pharmaceuticals, and cosmetics. Current performance evaluations of CDPs predominantly rely on empirical setups using simplistic metrics like Hamming distances or Pearson correlation. These methods are often inadequate due to their sensitivity to distortions, degradation, and their limitations to stationary statistics of printing and imaging. Additionally, machine learning-based approaches suffer from distribution biases and fail to generalize to unseen counterfeit samples. Given the critical importance of CDPs in preventing counterfeiting, including the counterfeit vaccines issue highlighted during the COVID-19 pandemic, there is an urgent need for provable performance guarantees across various criteria. This paper aims to establish a theoretical framework to derive optimal criteria for the analysis, optimization, and future development of CDP authentication technologies, ensuring their reliability and effectiveness in diverse security scenarios.
Authors: Shin'ya Yamaguchi, Kosuke Nishida
Abstract: Recent concept-based interpretable models have succeeded in providing meaningful explanations by pre-defined concept sets. However, the dependency on the pre-defined concepts restricts the application because of the limited number of concepts for explanations. This paper proposes a novel interpretable deep neural network called explanation bottleneck models (XBMs). XBMs generate a text explanation from the input without pre-defined concepts and then predict a final task prediction based on the generated explanation by leveraging pre-trained vision-language encoder-decoder models. To achieve both the target task performance and the explanation quality, we train XBMs through the target task loss with the regularization penalizing the explanation decoder via the distillation from the frozen pre-trained decoder. Our experiments, including a comparison to state-of-the-art concept bottleneck models, confirm that XBMs provide accurate and fluent natural language explanations without pre-defined concept sets. Code will be available at https://github.com/yshinya6/xbm/.
Authors: Nghia Nguyen, Minh Nhat Vu, Tung D. Ta, Baoru Huang, Thieu Vo, Ngan Le, Anh Nguyen
Abstract: Vision language models have played a key role in extracting meaningful features for various robotic applications. Among these, Contrastive Language-Image Pretraining (CLIP) is widely used in robotic tasks that require both vision and natural language understanding. However, CLIP was trained solely on static images paired with text prompts and has not yet been fully adapted for robotic tasks involving dynamic actions. In this paper, we introduce Robotic-CLIP to enhance robotic perception capabilities. We first gather and label large-scale action data, and then build our Robotic-CLIP by fine-tuning CLIP on 309,433 videos (~7.4 million frames) of action data using contrastive learning. By leveraging action data, Robotic-CLIP inherits CLIP's strong image performance while gaining the ability to understand actions in robotic contexts. Intensive experiments show that our Robotic-CLIP outperforms other CLIP-based models across various language-driven robotic tasks. Additionally, we demonstrate the practical effectiveness of Robotic-CLIP in real-world grasping applications.
Authors: Zhongxin Yu, Liang Chen, Zhiyun Zeng, Kunping Yang, Shaofei Luo, Shaorui Chen, Cheng Zhong
Abstract: Capturing different intensity and directions of light rays at the same scene Light field (LF) can encode the 3D scene cues into a 4D LF image which has a wide range of applications (i.e. post-capture refocusing and depth sensing). LF image super-resolution (SR) aims to improve the image resolution limited by the performance of LF camera sensor. Although existing methods have achieved promising results the practical application of these models is limited because they are not lightweight enough. In this paper we propose a lightweight model named LGFN which integrates the local and global features of different views and the features of different channels for LF image SR. Specifically owing to neighboring regions of the same pixel position in different sub-aperture images exhibit similar structural relationships we design a lightweight CNN-based feature extraction module (namely DGCE) to extract local features better through feature modulation. Meanwhile as the position beyond the boundaries in the LF image presents a large disparity we propose an efficient spatial attention module (namely ESAM) which uses decomposable large-kernel convolution to obtain an enlarged receptive field and an efficient channel attention module (namely ECAM). Compared with the existing LF image SR models with large parameter our model has a parameter of 0.45M and a FLOPs of 19.33G which has achieved a competitive effect. Extensive experiments with ablation studies demonstrate the effectiveness of our proposed method which ranked the second place in the Track 2 Fidelity & Efficiency of NTIRE2024 Light Field Super Resolution Challenge and the seventh place in the Track 1 Fidelity.
Authors: Harsh Yadav, Maximilian Schaefer, Kun Zhao, Tobias Meisen
Abstract: Motion prediction is an important aspect for Autonomous Driving (AD) and Advance Driver Assistance Systems (ADAS). Current state-of-the-art motion prediction methods rely on High Definition (HD) maps for capturing the surrounding context of the ego vehicle. Such systems lack scalability in real-world deployment as HD maps are expensive to produce and update in real-time. To overcome this issue, we propose Context Aware Scene Prediction Transformer (CASPFormer), which can perform multi-modal motion prediction from rasterized Bird-Eye-View (BEV) images. Our system can be integrated with any upstream perception module that is capable of generating BEV images. Moreover, CASPFormer directly decodes vectorized trajectories without any postprocessing. Trajectories are decoded recurrently using deformable attention, as it is computationally efficient and provides the network with the ability to focus its attention on the important spatial locations of the BEV images. In addition, we also address the issue of mode collapse for generating multiple scene-consistent trajectories by incorporating learnable mode queries. We evaluate our model on the nuScenes dataset and show that it reaches state-of-the-art across multiple metrics
Authors: Sophia Dowlatabadi, Bernhard Preim, Monique Meuschke
Abstract: In various fields, including medicine, age distributions are crucial. Despite widespread media coverage of health topics, there remains a need to enhance health communication. Narrative medical visualization is promising for improving information comprehension and retention. This study explores the most effective ways to present age distributions of diseases through narrative visualizations. We conducted a thorough analysis of existing visualizations, held workshops with a broad audience, and reviewed relevant literature. From this, we identified design choices focusing on comprehension, aesthetics, engagement, and memorability. We specifically tested three pictogram variants: pictograms as bars, stacked pictograms, and annotations. After evaluating 18 visualizations with 72 participants and three expert reviews, we determined that annotations were most effective for comprehension and aesthetics. However, traditional bar charts were preferred for engagement, and other variants were more memorable. The study provides a set of design recommendations based on these insights.
Authors: Amita Kamath, Cheng-Yu Hsieh, Kai-Wei Chang, Ranjay Krishna
Abstract: Several benchmarks have concluded that our best vision-language models (e.g., CLIP) are lacking in compositionality. Given an image, these benchmarks probe a model's ability to identify its associated caption amongst a set of compositional distractors. In response, a surge of recent proposals show improvements by finetuning CLIP with distractors as hard negatives. Our investigations reveal that these improvements have, in fact, been significantly overstated -- because existing benchmarks do not probe whether finetuned vision-language models remain invariant to hard positives. By curating an evaluation dataset with 112,382 hard negatives and hard positives, we uncover that including hard positives decreases CLIP's performance by 12.9%, while humans perform effortlessly at 99%. CLIP finetuned with hard negatives results in an even larger decrease, up to 38.7%. With this finding, we then produce a 1,775,259 image-text training set with both hard negative and hard positive captions. By training with both, we see improvements on existing benchmarks while simultaneously improving performance on hard positives, indicating a more robust improvement in compositionality. Our work suggests the need for future research to rigorously test and improve CLIP's understanding of semantic relationships between related "positive" concepts.
Authors: Xin Cai, Zhiyuan You, Hailong Zhang, Wentao Liu, Jinwei Gu, Tianfan Xue
Abstract: Lensless cameras offer significant advantages in size, weight, and cost compared to traditional lens-based systems. Without a focusing lens, lensless cameras rely on computational algorithms to recover the scenes from multiplexed measurements. However, current algorithms struggle with inaccurate forward imaging models and insufficient priors to reconstruct high-quality images. To overcome these limitations, we introduce a novel two-stage approach for consistent and photorealistic lensless image reconstruction. The first stage of our approach ensures data consistency by focusing on accurately reconstructing the low-frequency content with a spatially varying deconvolution method that adjusts to changes in the Point Spread Function (PSF) across the camera's field of view. The second stage enhances photorealism by incorporating a generative prior from pre-trained diffusion models. By conditioning on the low-frequency content retrieved in the first stage, the diffusion model effectively reconstructs the high-frequency details that are typically lost in the lensless imaging process, while also maintaining image fidelity. Our method achieves a superior balance between data fidelity and visual quality compared to existing methods, as demonstrated with two popular lensless systems, PhlatCam and DiffuserCam. Project website: https://phocolens.github.io/.
Authors: Xin Li, Siyuan Huang, Qiaojun Yu, Zhengkai Jiang, Ce Hao, Yimeng Zhu, Hongsheng Li, Peng Gao, Cewu Lu
Abstract: Automating garment manipulation poses a significant challenge for assistive robotics due to the diverse and deformable nature of garments. Traditional approaches typically require separate models for each garment type, which limits scalability and adaptability. In contrast, this paper presents a unified approach using vision-language models (VLMs) to improve keypoint prediction across various garment categories. By interpreting both visual and semantic information, our model enables robots to manage different garment states with a single model. We created a large-scale synthetic dataset using advanced simulation techniques, allowing scalable training without extensive real-world data. Experimental results indicate that the VLM-based method significantly enhances keypoint detection accuracy and task success rates, providing a more flexible and general solution for robotic garment manipulation. In addition, this research also underscores the potential of VLMs to unify various garment manipulation tasks within a single framework, paving the way for broader applications in home automation and assistive robotics for future.
Authors: Theo Larcher, Lukas Picek, Benjamin Deneu, Titouan Lorieul, Maximilien Servajean, Alexis Joly
Abstract: This paper describes a deep-SDM framework, MALPOLON. Written in Python and built upon the PyTorch library, this framework aims to facilitate training and inferences of deep species distribution models (deep-SDM) and sharing for users with only general Python language skills (e.g., modeling ecologists) who are interested in testing deep learning approaches to build new SDMs. More advanced users can also benefit from the framework's modularity to run more specific experiments by overriding existing classes while taking advantage of press-button examples to train neural networks on multiple classification tasks using custom or provided raw and pre-processed datasets. The framework is open-sourced on GitHub and PyPi along with extensive documentation and examples of use in various scenarios. MALPOLON offers straightforward installation, YAML-based configuration, parallel computing, multi-GPU utilization, baseline and foundational models for benchmarking, and extensive tutorials/documentation, aiming to enhance accessibility and performance scalability for ecologists and researchers.
Authors: Fernando Cladera, Kenneth Chaney, M. Ani Hsieh, Camillo J. Taylor, Vijay Kumar
Abstract: Traditionally, unmanned aerial vehicles (UAVs) rely on CMOS-based cameras to collect images about the world below. One of the most successful applications of UAVs is to generate orthomosaics or orthomaps, in which a series of images are integrated together to develop a larger map. However, the use of CMOS-based cameras with global or rolling shutters mean that orthomaps are vulnerable to challenging light conditions, motion blur, and high-speed motion of independently moving objects under the camera. Event cameras are less sensitive to these issues, as their pixels are able to trigger asynchronously on brightness changes. This work introduces the first orthomosaic approach using event cameras. In contrast to existing methods relying only on CMOS cameras, our approach enables map generation even in challenging light conditions, including direct sunlight and after sunset.
Authors: Justin Kerr, Chung Min Kim, Mingxuan Wu, Brent Yi, Qianqian Wang, Ken Goldberg, Angjoo Kanazawa
Abstract: Humans can learn to manipulate new objects by simply watching others; providing robots with the ability to learn from such demonstrations would enable a natural interface specifying new behaviors. This work develops Robot See Robot Do (RSRD), a method for imitating articulated object manipulation from a single monocular RGB human demonstration given a single static multi-view object scan. We first propose 4D Differentiable Part Models (4D-DPM), a method for recovering 3D part motion from a monocular video with differentiable rendering. This analysis-by-synthesis approach uses part-centric feature fields in an iterative optimization which enables the use of geometric regularizers to recover 3D motions from only a single video. Given this 4D reconstruction, the robot replicates object trajectories by planning bimanual arm motions that induce the demonstrated object part motion. By representing demonstrations as part-centric trajectories, RSRD focuses on replicating the demonstration's intended behavior while considering the robot's own morphological limits, rather than attempting to reproduce the hand's motion. We evaluate 4D-DPM's 3D tracking accuracy on ground truth annotated 3D part trajectories and RSRD's physical execution performance on 9 objects across 10 trials each on a bimanual YuMi robot. Each phase of RSRD achieves an average of 87% success rate, for a total end-to-end success rate of 60% across 90 trials. Notably, this is accomplished using only feature fields distilled from large pretrained vision models -- without any task-specific training, fine-tuning, dataset collection, or annotation. Project page: https://robot-see-robot-do.github.io
Authors: Xin Ma, Guorong Wu, Seong Jae Hwang, Won Hwa Kim
Abstract: Tremendous recent literature show that associations between different brain regions, i.e., brain connectivity, provide early symptoms of neurological disorders. Despite significant efforts made for graph neural network (GNN) techniques, their focus on graph nodes makes the state-of-the-art GNN methods not suitable for classifying brain connectivity as graphs where the objective is to characterize disease-relevant network dysfunction patterns on graph links. To address this issue, we propose Multi-resolution Edge Network (MENET) to detect disease-specific connectomic benchmarks with high discrimination power across diagnostic categories. The core of MENET is a novel graph edge-wise transform that we propose, which allows us to capture multi-resolution ``connectomic'' features. Using a rich set of the connectomic features, we devise a graph learning framework to jointly select discriminative edges and assign diagnostic labels for graphs. Experiments on two real datasets show that MENET accurately predicts diagnostic labels and identify brain connectivities highly associated with neurological disorders such as Alzheimer's Disease and Attention-Deficit/Hyperactivity Disorder.
Authors: Mohammad Abuzar Hashemi, Zhanghexuan Li, Mihir Chauhan, Yan Shen, Abhishek Satbhai, Mir Basheer Ali, Mingchen Gao, Sargur Srihari
Abstract: Pre-training visual and textual representations from large-scale image-text pairs is becoming a standard approach for many downstream vision-language tasks. The transformer-based models learn inter and intra-modal attention through a list of self-supervised learning tasks. This paper proposes LAViTeR, a novel architecture for visual and textual representation learning. The main module, Visual Textual Alignment (VTA) will be assisted by two auxiliary tasks, GAN-based image synthesis and Image Captioning. We also propose a new evaluation metric measuring the similarity between the learnt visual and textual embedding. The experimental results on two public datasets, CUB and MS-COCO, demonstrate superior visual and textual representation alignment in the joint feature embedding space
Authors: Jiahao Qin, Yitao Xu, Zong Lu, Xiaojun Zhang
Abstract: In the realm of multimodal data integration, feature alignment plays a pivotal role. This paper introduces an innovative approach to feature alignment that revolutionizes the fusion of multimodal information. Our method employs a novel iterative process of telescopic displacement and expansion of feature representations across different modalities, culminating in a coherent unified representation within a shared feature space. This sophisticated technique demonstrates a remarkable ability to capture and leverage complex crossmodal interactions at the highest levels of abstraction. As a result, we observe significant enhancements in the performance of multimodal learning tasks. Through rigorous comparative analysis, we establish the superiority of our approach over existing multimodal fusion paradigms across a diverse array of applications. Comprehensive empirical evaluations conducted on multifaceted datasets encompassing temporal sequences, visual data, and textual information provide compelling evidence that our method achieves unprecedented benchmarks in the field. This work not only advances the state of the art in multimodal learning but also opens new avenues for exploring the synergies between disparate data modalities in complex analytical scenarios.
Authors: Natalia Frumkin, Dibakar Gope, Diana Marculescu
Abstract: Quantization scale and bit-width are the most important parameters when considering how to quantize a neural network. Prior work focuses on optimizing quantization scales in a global manner through gradient methods (gradient descent \& Hessian analysis). Yet, when applying perturbations to quantization scales, we observe a very jagged, highly non-smooth test loss landscape. In fact, small perturbations in quantization scale can greatly affect accuracy, yielding a $0.5-0.8\%$ accuracy boost in 4-bit quantized vision transformers (ViTs). In this regime, gradient methods break down, since they cannot reliably reach local minima. In our work, dubbed Evol-Q, we use evolutionary search to effectively traverse the non-smooth landscape. Additionally, we propose using an infoNCE loss, which not only helps combat overfitting on the small calibration dataset ($1,000$ images) but also makes traversing such a highly non-smooth surface easier. Evol-Q improves the top-1 accuracy of a fully quantized ViT-Base by $10.30\%$, $0.78\%$, and $0.15\%$ for $3$-bit, $4$-bit, and $8$-bit weight quantization levels. Extensive experiments on a variety of CNN and ViT architectures further demonstrate its robustness in extreme quantization scenarios. Our code is available at https://github.com/enyac-group/evol-q
Authors: Ofir Gordon, Elad Cohen, Hai Victor Habi, Arnon Netzer
Abstract: Quantization is a key method for deploying deep neural networks on edge devices with limited memory and computation resources. Recent improvements in Post-Training Quantization (PTQ) methods were achieved by an additional local optimization process for learning the weight quantization rounding policy. However, a gap exists when employing network-wise optimization with small representative datasets. In this paper, we propose a new method for enhanced PTQ (EPTQ) that employs a network-wise quantization optimization process, which benefits from considering cross-layer dependencies during optimization. EPTQ enables network-wise optimization with a small representative dataset using a novel sample-layer attention score based on a label-free Hessian matrix upper bound. The label-free approach makes our method suitable for the PTQ scheme. We give a theoretical analysis for the said bound and use it to construct a knowledge distillation loss that guides the optimization to focus on the more sensitive layers and samples. In addition, we leverage the Hessian upper bound to improve the weight quantization parameters selection by focusing on the more sensitive elements in the weight tensors. Empirically, by employing EPTQ we achieve state-of-the-art results on various models, tasks, and datasets, including ImageNet classification, COCO object detection, and Pascal-VOC for semantic segmentation.
Authors: Xiaoting Yin, Hao Shi, Jiaan Chen, Ze Wang, Yaozu Ye, Kailun Yang, Kaiwei Wang
Abstract: Human pose estimation is a fundamental and appealing task in computer vision. Although traditional cameras are commonly applied, their reliability decreases in scenarios under high dynamic range or heavy motion blur, where event cameras offer a robust solution. Predominant event-based methods accumulate events into frames, ignoring the asynchronous and high temporal resolution that is crucial for distinguishing distinct actions. To address this issue and to unlock the 3D potential of event information, we introduce two 3D event representations: the Rasterized Event Point Cloud (RasEPC) and the Decoupled Event Voxel (DEV). The RasEPC aggregates events within concise temporal slices at identical positions, preserving their 3D attributes along with statistical information, thereby significantly reducing memory and computational demands. Meanwhile, the DEV representation discretizes events into voxels and projects them across three orthogonal planes, utilizing decoupled event attention to retrieve 3D cues from the 2D planes. Furthermore, we develop and release EV-3DPW, a synthetic event-based dataset crafted to facilitate training and quantitative analysis in outdoor scenes. Our methods are tested on the DHP19 public dataset, MMHPSD dataset, and our EV-3DPW dataset, with further qualitative validation via a derived driving scene dataset EV-JAAD and an outdoor collection vehicle. Our code and dataset have been made publicly available at https://github.com/MasterHow/EventPointPose.
Authors: Kristen Grauman, Andrew Westbury, Lorenzo Torresani, Kris Kitani, Jitendra Malik, Triantafyllos Afouras, Kumar Ashutosh, Vijay Baiyya, Siddhant Bansal, Bikram Boote, Eugene Byrne, Zach Chavis, Joya Chen, Feng Cheng, Fu-Jen Chu, Sean Crane, Avijit Dasgupta, Jing Dong, Maria Escobar, Cristhian Forigua, Abrham Gebreselasie, Sanjay Haresh, Jing Huang, Md Mohaiminul Islam, Suyog Jain, Rawal Khirodkar, Devansh Kukreja, Kevin J Liang, Jia-Wei Liu, Sagnik Majumder, Yongsen Mao, Miguel Martin, Effrosyni Mavroudi, Tushar Nagarajan, Francesco Ragusa, Santhosh Kumar Ramakrishnan, Luigi Seminara, Arjun Somayazulu, Yale Song, Shan Su, Zihui Xue, Edward Zhang, Jinxu Zhang, Angela Castillo, Changan Chen, Xinzhu Fu, Ryosuke Furuta, Cristina Gonzalez, Prince Gupta, Jiabo Hu, Yifei Huang, Yiming Huang, Weslie Khoo, Anush Kumar, Robert Kuo, Sach Lakhavani, Miao Liu, Mi Luo, Zhengyi Luo, Brighid Meredith, Austin Miller, Oluwatumininu Oguntola, Xiaqing Pan, Penny Peng, Shraman Pramanick, Merey Ramazanova, Fiona Ryan, Wei Shan, Kiran Somasundaram, Chenan Song, Audrey Southerland, Masatoshi Tateno, Huiyu Wang, Yuchen Wang, Takuma Yagi, Mingfei Yan, Xitong Yang, Zecheng Yu, Shengxin Cindy Zha, Chen Zhao, Ziwei Zhao, Zhifan Zhu, Jeff Zhuo, Pablo Arbelaez, Gedas Bertasius, David Crandall, Dima Damen, Jakob Engel, Giovanni Maria Farinella, Antonino Furnari, Bernard Ghanem, Judy Hoffman, C. V. Jawahar, Richard Newcombe, Hyun Soo Park, James M. Rehg, Yoichi Sato, Manolis Savva, Jianbo Shi, Mike Zheng Shou, Michael Wray
Abstract: We present Ego-Exo4D, a diverse, large-scale multimodal multiview video dataset and benchmark challenge. Ego-Exo4D centers around simultaneously-captured egocentric and exocentric video of skilled human activities (e.g., sports, music, dance, bike repair). 740 participants from 13 cities worldwide performed these activities in 123 different natural scene contexts, yielding long-form captures from 1 to 42 minutes each and 1,286 hours of video combined. The multimodal nature of the dataset is unprecedented: the video is accompanied by multichannel audio, eye gaze, 3D point clouds, camera poses, IMU, and multiple paired language descriptions -- including a novel "expert commentary" done by coaches and teachers and tailored to the skilled-activity domain. To push the frontier of first-person video understanding of skilled human activity, we also present a suite of benchmark tasks and their annotations, including fine-grained activity understanding, proficiency estimation, cross-view translation, and 3D hand/body pose. All resources are open sourced to fuel new research in the community. Project page: http://ego-exo4d-data.org/
Authors: Zhiyu Pan, Yongjie Duan, Jianjiang Feng, Jie Zhou
Abstract: In fingerprint matching, fixed-length descriptors generally offer greater efficiency compared to minutiae set, but the recognition accuracy is not as good as that of the latter. Although much progress has been made in deep learning based fixed-length descriptors recently, they often fall short when dealing with incomplete or partial fingerprints, diverse fingerprint poses, and significant background noise. In this paper, we propose a three-dimensional representation called Fixed-length Dense Descriptor (FDD) for efficient fingerprint matching. FDD features great spatial properties, enabling it to capture the spatial relationships of the original fingerprints, thereby enhancing interpretability and robustness. Our experiments on various fingerprint datasets reveal that FDD outperforms other fixed-length descriptors, especially in matching fingerprints of different areas, cross-modal fingerprint matching, and fingerprint matching with background noise.
Authors: Sharath Girish, Kamal Gupta, Abhinav Shrivastava
Abstract: Recently, 3D Gaussian splatting (3D-GS) has gained popularity in novel-view scene synthesis. It addresses the challenges of lengthy training times and slow rendering speeds associated with Neural Radiance Fields (NeRFs). Through rapid, differentiable rasterization of 3D Gaussians, 3D-GS achieves real-time rendering and accelerated training. They, however, demand substantial memory resources for both training and storage, as they require millions of Gaussians in their point cloud representation for each scene. We present a technique utilizing quantized embeddings to significantly reduce per-point memory storage requirements and a coarse-to-fine training strategy for a faster and more stable optimization of the Gaussian point clouds. Our approach develops a pruning stage which results in scene representations with fewer Gaussians, leading to faster training times and rendering speeds for real-time rendering of high resolution scenes. We reduce storage memory by more than an order of magnitude all while preserving the reconstruction quality. We validate the effectiveness of our approach on a variety of datasets and scenes preserving the visual quality while consuming 10-20x lesser memory and faster training/inference speed. Project page and code is available https://efficientgaussian.github.io
Authors: Zigeng Chen, Gongfan Fang, Xinyin Ma, Xinchao Wang
Abstract: Current approaches for compressing the Segment Anything Model (SAM) yield commendable results, yet necessitate extensive data to train a new network from scratch. Employing conventional pruning techniques can remarkably reduce data requirements but would suffer from a degradation in performance. To address this challenging trade-off, we introduce SlimSAM, a novel data-efficient SAM compression method that achieves superior performance with extremely less training data. The essence of SlimSAM is encapsulated in the alternate slimming framework which effectively enhances knowledge inheritance under severely limited training data availability and exceptional pruning ratio. Diverging from prior techniques, our framework progressively compresses the model by alternately pruning and distilling distinct, decoupled sub-structures. Disturbed Taylor pruning is also proposed to address the misalignment between the pruning objective and training target, thereby boosting the post-distillation after pruning. SlimSAM yields significant performance improvements while demanding over 10 times less training data than any other existing compression methods. Even when compared to the original SAM, SlimSAM achieves approaching performance while reducing parameter counts to merely 1.4% (9.1M), MACs to 0.8% (23G), and requiring only 0.1% (10k) of the SAM training data. The code is available at http://github.com/czg1225/SlimSAM.
Authors: Jionghao Wang, Yuan Liu, Zhiyang Dou, Zhengming Yu, Yongqing Liang, Cheng Lin, Xin Li, Wenping Wang, Rong Xie, Li Song
Abstract: In this paper, we introduce a novel text-to-avatar generation method that separately generates the human body and the clothes and allows high-quality animation on the generated avatar. While recent advancements in text-to-avatar generation have yielded diverse human avatars from text prompts, these methods typically combine all elements-clothes, hair, and body-into a single 3D representation. Such an entangled approach poses challenges for downstream tasks like editing or animation. To overcome these limitations, we propose a novel disentangled 3D avatar representation named Sequentially Offset-SMPL (SO-SMPL), building upon the SMPL model. SO-SMPL represents the human body and clothes with two separate meshes but associates them with offsets to ensure the physical alignment between the body and the clothes. Then, we design a Score Distillation Sampling (SDS)-based distillation framework to generate the proposed SO-SMPL representation from text prompts. Our approach not only achieves higher texture and geometry quality and better semantic alignment with text prompts, but also significantly improves the visual quality of character animation, virtual try-on, and avatar editing. Project page: https://shanemankiw.github.io/SO-SMPL/.
Authors: Haifeng Huang, Yilun Chen, Zehan Wang, Rongjie Huang, Runsen Xu, Tai Wang, Luping Liu, Xize Cheng, Yang Zhao, Jiangmiao Pang, Zhou Zhao
Abstract: Recent advancements in 3D Large Language Models (LLMs) have demonstrated promising capabilities for 3D scene understanding. However, previous methods exhibit deficiencies in general referencing and grounding capabilities for intricate scene comprehension. In this paper, we introduce the use of object identifiers and object-centric representations to interact with scenes at the object level. Specifically, we decompose the input 3D scene into a set of object proposals, each assigned a unique identifier token, which enables efficient object referencing and grounding during user-assistant interactions. Given the scarcity of scene-language data, we model the scene embeddings as a sequence of explicit object-level embeddings, derived from semantic-rich 2D or 3D representations. By employing object identifiers, we transform diverse 3D scene-language tasks into a unified question-answering format, facilitating joint training without the need for additional task-specific heads. With minimal fine-tuning on all downstream tasks, our model significantly outperforms existing methods on benchmarks including ScanRefer, Multi3DRefer, Scan2Cap, ScanQA, and SQA3D.
Authors: Mirela Ostrek, Carol O'Sullivan, Michael J. Black, Justus Thies
Abstract: We present ESP, a novel method for context-aware full-body generation, that enables photo-realistic synthesis and inpainting of people wearing clothing that is semantically appropriate for the scene depicted in an input photograph. ESP is conditioned on a 2D pose and contextual cues that are extracted from the photograph of the scene and integrated into the generation process, where the clothing is modeled explicitly with human parsing masks (HPM). Generated HPMs are used as tight guiding masks for inpainting, such that no changes are made to the original background. Our models are trained on a dataset containing a set of in-the-wild photographs of people covering a wide range of different environments. The method is analyzed quantitatively and qualitatively, and we show that ESP outperforms the state-of-the-art on the task of contextual full-body generation.
Authors: Kaichen Zhou, Jia-Wang Bian, Qian Xie, Jian-Qing Zheng, Niki Trigoni, Andrew Markham
Abstract: Despite advancements in self-supervised monocular depth estimation, challenges persist in dynamic scenarios due to the dependence on assumptions about a static world. In this paper, we present Manydepth2, a Motion-Guided Cost Volume Depth Net, to achieve precise depth estimation for both dynamic objects and static backgrounds, all while maintaining computational efficiency. To tackle the challenges posed by dynamic content, we incorporate optical flow and coarse monocular depth to create a novel static reference frame. This frame is then utilized to build a motion-guided cost volume in collaboration with the target frame. Additionally, to enhance the accuracy and resilience of the network structure, we introduce an attention-based depth net architecture to effectively integrate information from feature maps with varying resolutions. Compared to methods with similar computational costs, Manydepth2 achieves a significant reduction of approximately five percent in root-mean-square error for self-supervised monocular depth estimation on the KITTI-2015 dataset. The code could be found: https://github.com/kaichen-z/Manydepth2
Authors: Rosco Hunter, {\L}ukasz Dudziak, Mohamed S. Abdelfattah, Abhinav Mehrotra, Sourav Bhattacharya, Hongkai Wen
Abstract: Text-to-image diffusion models have demonstrated unprecedented capabilities for flexible and realistic image synthesis. Nevertheless, these models rely on a time-consuming sampling procedure, which has motivated attempts to reduce their latency. When improving efficiency, researchers often use the original diffusion model to train an additional network designed specifically for fast image generation. In contrast, our approach seeks to reduce latency directly, without any retraining, fine-tuning, or knowledge distillation. In particular, we find the repeated calculation of attention maps to be costly yet redundant, and instead suggest reusing them during sampling. Our specific reuse strategies are based on ODE theory, which implies that the later a map is reused, the smaller the distortion in the final image. We empirically compare these reuse strategies with few-step sampling procedures of comparable latency, finding that reuse generates images that are closer to those produced by the original high-latency diffusion model.
Authors: Xuewen Liu, Zhikai Li, Junrui Xiao, Qingyi Gu
Abstract: Diffusion models have achieved great success in image generation tasks through iterative noise estimation. However, the heavy denoising process and complex neural networks hinder their low-latency applications in real-world scenarios. Quantization can effectively reduce model complexity, and post-training quantization (PTQ), which does not require fine-tuning, is highly promising for compressing and accelerating diffusion models. Unfortunately, we find that due to the highly dynamic distribution of activations in different denoising steps, existing PTQ methods for diffusion models suffer from distribution mismatch issues at both calibration sample level and reconstruction output level, which makes the performance far from satisfactory, especially in low-bit cases. In this paper, we propose Enhanced Distribution Alignment for Post-Training Quantization of Diffusion Models (EDA-DM) to address the above issues. Specifically, at the calibration sample level, we select calibration samples based on the density and variety in the latent space, thus facilitating the alignment of their distribution with the overall samples; and at the reconstruction output level, we modify the loss of block reconstruction with the losses of layers, aligning the outputs of quantized model and full-precision model at different network granularity. Extensive experiments demonstrate that EDA-DM significantly outperforms the existing PTQ methods across various models (DDIM, LDM-4, LDM-8, Stable-Diffusion) and different datasets (CIFAR-10, LSUN-Bedroom, LSUN-Church, ImageNet, MS-COCO).
Authors: Marcos V. Conde, Gregor Geigle, Radu Timofte
Abstract: Image restoration is a fundamental problem that involves recovering a high-quality clean image from its degraded observation. All-In-One image restoration models can effectively restore images from various types and levels of degradation using degradation-specific information as prompts to guide the restoration model. In this work, we present the first approach that uses human-written instructions to guide the image restoration model. Given natural language prompts, our model can recover high-quality images from their degraded counterparts, considering multiple degradation types. Our method, InstructIR, achieves state-of-the-art results on several restoration tasks including image denoising, deraining, deblurring, dehazing, and (low-light) image enhancement. InstructIR improves +1dB over previous all-in-one restoration methods. Moreover, our dataset and results represent a novel benchmark for new research on text-guided image restoration and enhancement. Our code, datasets and models are available at: https://github.com/mv-lab/InstructIR
Authors: Wenjun Hou, Yi Cheng, Kaishuai Xu, Yan Hu, Wenjie Li, Jiang Liu
Abstract: Previous research on radiology report generation has made significant progress in terms of increasing the clinical accuracy of generated reports. In this paper, we emphasize another crucial quality that it should possess, i.e., inter-report consistency, which refers to the capability of generating consistent reports for semantically equivalent radiographs. This quality is even of greater significance than the overall report accuracy in terms of ensuring the system's credibility, as a system prone to providing conflicting results would severely erode users' trust. Regrettably, existing approaches struggle to maintain inter-report consistency, exhibiting biases towards common patterns and susceptibility to lesion variants. To address this issue, we propose ICON, which improves the inter-report consistency of radiology report generation. Aiming to enhance the system's ability to capture similarities in semantically equivalent lesions, our approach first involves extracting lesions from input images and examining their characteristics. Then, we introduce a lesion-aware mixup technique to ensure that the representations of the semantically equivalent lesions align with the same attributes, achieved through a linear combination during the training phase. Extensive experiments on three publicly available chest X-ray datasets verify the effectiveness of our approach, both in terms of improving the consistency and accuracy of the generated reports.
Authors: Bin Li, Ye Shi, Qian Yu, Jingya Wang
Abstract: Unsupervised cross-domain image retrieval (UCIR) aims to retrieve images sharing the same category across diverse domains without relying on labeled data. Prior approaches have typically decomposed the UCIR problem into two distinct tasks: intra-domain representation learning and cross-domain feature alignment. However, these segregated strategies overlook the potential synergies between these tasks. This paper introduces ProtoOT, a novel Optimal Transport formulation explicitly tailored for UCIR, which integrates intra-domain feature representation learning and cross-domain alignment into a unified framework. ProtoOT leverages the strengths of the K-means clustering method to effectively manage distribution imbalances inherent in UCIR. By utilizing K-means for generating initial prototypes and approximating class marginal distributions, we modify the constraints in Optimal Transport accordingly, significantly enhancing its performance in UCIR scenarios. Furthermore, we incorporate contrastive learning into the ProtoOT framework to further improve representation learning. This encourages local semantic consistency among features with similar semantics, while also explicitly enforcing separation between features and unmatched prototypes, thereby enhancing global discriminativeness. ProtoOT surpasses existing state-of-the-art methods by a notable margin across benchmark datasets. Notably, on DomainNet, ProtoOT achieves an average P@200 enhancement of 18.17%, and on Office-Home, it demonstrates a P@15 improvement of 3.83%.
Authors: Zunnan Xu, Yukang Lin, Haonan Han, Sicheng Yang, Ronghui Li, Yachao Zhang, Xiu Li
Abstract: Gesture synthesis is a vital realm of human-computer interaction, with wide-ranging applications across various fields like film, robotics, and virtual reality. Recent advancements have utilized the diffusion model and attention mechanisms to improve gesture synthesis. However, due to the high computational complexity of these techniques, generating long and diverse sequences with low latency remains a challenge. We explore the potential of state space models (SSMs) to address the challenge, implementing a two-stage modeling strategy with discrete motion priors to enhance the quality of gestures. Leveraging the foundational Mamba block, we introduce MambaTalk, enhancing gesture diversity and rhythm through multimodal integration. Extensive experiments demonstrate that our method matches or exceeds the performance of state-of-the-art models.
Authors: Abdelrahman Shaker, Syed Talal Wasim, Martin Danelljan, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan
Abstract: Recently, transformer-based approaches have shown promising results for semi-supervised video object segmentation. However, these approaches typically struggle on long videos due to increased GPU memory demands, as they frequently expand the memory bank every few frames. We propose a transformer-based approach, named MAVOS, that introduces an optimized and dynamic long-term modulated cross-attention (MCA) memory to model temporal smoothness without requiring frequent memory expansion. The proposed MCA effectively encodes both local and global features at various levels of granularity while efficiently maintaining consistent speed regardless of the video length. Extensive experiments on multiple benchmarks, LVOS, Long-Time Video, and DAVIS 2017, demonstrate the effectiveness of our proposed contributions leading to real-time inference and markedly reduced memory demands without any degradation in segmentation accuracy on long videos. Compared to the best existing transformer-based approach, our MAVOS increases the speed by 7.6x, while significantly reducing the GPU memory by 87% with comparable segmentation performance on short and long video datasets. Notably on the LVOS dataset, our MAVOS achieves a J&F score of 63.3% while operating at 37 frames per second (FPS) on a single V100 GPU. Our code and models will be publicly available at: https://github.com/Amshaker/MAVOS.
Authors: Zheling Meng, Bo Peng, Jing Dong
Abstract: Watermarking is a tool for actively identifying and attributing the images generated by latent diffusion models. Existing methods face the dilemma of image quality and watermark robustness. Watermarks with superior image quality usually have inferior robustness against attacks such as blurring and JPEG compression, while watermarks with superior robustness usually significantly damage image quality. This dilemma stems from the traditional paradigm where watermarks are injected and detected in pixel space, relying on pixel perturbation for watermark detection and resilience against attacks. In this paper, we highlight that an effective solution to the problem is to both inject and detect watermarks in the latent diffusion space, and propose Latent Watermark with a progressive training strategy. It weakens the direct connection between quality and robustness and thus alleviates their contradiction. We conduct evaluations on two datasets and against 10 watermark attacks. Six metrics measure the image quality and watermark robustness. Results show that compared to the recently proposed methods such as StableSignature, StegaStamp, RoSteALS, LaWa, TreeRing, and DiffuseTrace, LW not only surpasses them in terms of robustness but also offers superior image quality. Our code will be available at https://github.com/RichardSunnyMeng/LatentWatermark.
Authors: Bonan Liu, Guoyang Zhao, Jianhao Jiao, Guang Cai, Chengyang Li, Handi Yin, Yuyang Wang, Ming Liu, Pan Hui
Abstract: A Colored point cloud, as a simple and efficient 3D representation, has many advantages in various fields, including robotic navigation and scene reconstruction. This representation is now commonly used in 3D reconstruction tasks relying on cameras and LiDARs. However, fusing data from these two types of sensors is poorly performed in many existing frameworks, leading to unsatisfactory mapping results, mainly due to inaccurate camera poses. This paper presents OmniColor, a novel and efficient algorithm to colorize point clouds using an independent 360-degree camera. Given a LiDAR-based point cloud and a sequence of panorama images with initial coarse camera poses, our objective is to jointly optimize the poses of all frames for mapping images onto geometric reconstructions. Our pipeline works in an off-the-shelf manner that does not require any feature extraction or matching process. Instead, we find optimal poses by directly maximizing the photometric consistency of LiDAR maps. In experiments, we show that our method can overcome the severe visual distortion of omnidirectional images and greatly benefit from the wide field of view (FOV) of 360-degree cameras to reconstruct various scenarios with accuracy and stability. The code will be released at https://github.com/liubonan123/OmniColor/.
Authors: Xinya Chen, Hanlei Guo, Yanrui Bin, Shangzhan Zhang, Yuanbo Yang, Yue Wang, Yujun Shen, Yiyi Liao
Abstract: Collecting accurate camera poses of training images has been shown to well serve the learning of 3D-aware generative adversarial networks (GANs) yet can be quite expensive in practice. This work targets learning 3D-aware GANs from unposed images, for which we propose to perform on-the-fly pose estimation of training images with a learned template feature field (TeFF). Concretely, in addition to a generative radiance field as in previous approaches, we ask the generator to also learn a field from 2D semantic features while sharing the density from the radiance field. Such a framework allows us to acquire a canonical 3D feature template leveraging the dataset mean discovered by the generative model, and further efficiently estimate the pose parameters on real data. Experimental results on various challenging datasets demonstrate the superiority of our approach over state-of-the-art alternatives from both the qualitative and the quantitative perspectives.
Authors: Mirza Tanzim Sami, Da Yan, Saugat Adhikari, Lyuheng Yuan, Jiao Han, Zhe Jiang, Jalal Khalil, Yang Zhou
Abstract: Accurate and timely mapping of flood extent from high-resolution satellite imagery plays a crucial role in disaster management such as damage assessment and relief activities. However, current state-of-the-art solutions are based on U-Net, which can-not segment the flood pixels accurately due to the ambiguous pixels (e.g., tree canopies, clouds) that prevent a direct judgement from only the spectral features. Thanks to the digital elevation model (DEM) data readily available from sources such as United States Geological Survey (USGS), this work explores the use of an elevation map to improve flood extent mapping. We propose, EvaNet, an elevation-guided segmentation model based on the encoder-decoder architecture with two novel techniques: (1) a loss function encoding the physical law of gravity that if a location is flooded (resp. dry), then its adjacent locations with a lower (resp. higher) elevation must also be flooded (resp. dry); (2) a new (de)convolution operation that integrates the elevation map by a location sensitive gating mechanism to regulate how much spectral features flow through adjacent layers. Extensive experiments show that EvaNet significantly outperforms the U-Net baselines, and works as a perfect drop-in replacement for U-Net in existing solutions to flood extent mapping.
Authors: Fangzhou Han, Lingyu Si, Hongwei Dong, Lamei Zhang, Hao Chen, Bo Du
Abstract: Artificial intelligence generative content (AIGC) has significantly impacted image generation in the field of remote sensing. However, the equally important area of remote sensing image (RSI) editing has not received sufficient attention. Deep learning based editing methods generally involve two sequential stages: generation and editing. During the generation stage, consistency in content and details between the original and edited images must be maintained, while in the editing stage, controllability and accuracy of the edits should be ensured. For natural images, these challenges can be tackled by training generative backbones on large-scale benchmark datasets and using text guidance based on vision-language models (VLMs). However, these previously effective approaches become less viable for RSIs due to two reasons: First, existing generative RSI benchmark datasets do not fully capture the diversity of remote sensing scenarios, particularly in terms of variations in sensors, object types, and resolutions. Consequently, the generalization capacity of the trained backbone model is often inadequate for universal editing tasks on RSIs. Second, the large spatial resolution of RSIs exacerbates the problem in VLMs where a single text semantic corresponds to multiple image semantics, leading to the introduction of incorrect semantics when using text to guide RSI editing. To solve above problems, this paper proposes a text-guided RSI editing method that is controllable but stable, and can be trained using only a single image. It adopts a multi-scale training approach to preserve consistency without the need for training on extensive benchmark datasets, while leveraging RSI pre-trained VLMs and prompt ensembling (PE) to ensure accuracy and controllability in the text-guided editing process.
Authors: Ancheng Lin, Jun Li
Abstract: Accurately reconstructing a 3D scene including explicit geometry information is both attractive and challenging. Geometry reconstruction can benefit from incorporating differentiable appearance models, such as Neural Radiance Fields and 3D Gaussian Splatting (3DGS). However, existing methods encounter efficiency issues due to indirect geometry learning and the paradigm of separately modeling geometry and surface appearance. In this work, we propose a learnable scene model that incorporates 3DGS with an explicit geometry representation, namely a mesh. Our model learns the mesh and appearance in an end-to-end manner, where we bind 3D Gaussians to the mesh faces and perform differentiable rendering of 3DGS to obtain photometric supervision. The model creates an effective information pathway to supervise the learning of both 3DGS and mesh. Experimental results demonstrate that the learned scene model not only achieves state-of-the-art efficiency and rendering quality but also supports manipulation using the explicit mesh. In addition, our model has a unique advantage in adapting to scene updates, thanks to the end-to-end learning of both mesh and appearance.
Authors: Daniel Bogdoll, Iramm Hamdard, Lukas Namgyu R\"o{\ss}ler, Felix Geisler, Muhammed Bayram, Felix Wang, Jan Imhof, Miguel de Campos, Anushervon Tabarov, Yitian Yang, Hanno Gottschalk, J. Marius Z\"ollner
Abstract: The scale-up of autonomous vehicles depends heavily on their ability to deal with anomalies, such as rare objects on the road. In order to handle such situations, it is necessary to detect anomalies in the first place. Anomaly detection for autonomous driving has made great progress in the past years but suffers from poorly designed benchmarks with a strong focus on camera data. In this work, we propose AnoVox, the largest benchmark for ANOmaly detection in autonomous driving to date. AnoVox incorporates large-scale multimodal sensor data and spatial VOXel ground truth, allowing for the comparison of methods independent of their used sensor. We propose a formal definition of normality and provide a compliant training dataset. AnoVox is the first benchmark to contain both content and temporal anomalies.
Authors: Mingshuang Luo, Ruibing Hou, Zhuo Li, Hong Chang, Zimo Liu, Yaowei Wang, Shiguang Shan
Abstract: This paper presents M$^3$GPT, an advanced $\textbf{M}$ultimodal, $\textbf{M}$ultitask framework for $\textbf{M}$otion comprehension and generation. M$^3$GPT operates on three fundamental principles. The first focuses on creating a unified representation space for various motion-relevant modalities. We employ discrete vector quantization for multimodal control and generation signals, such as text, music and motion/dance, enabling seamless integration into a large language model (LLM) with a single vocabulary. The second involves modeling model generation directly in the raw motion space. This strategy circumvents the information loss associated with discrete tokenizer, resulting in more detailed and comprehensive model generation. Third, M$^3$GPT learns to model the connections and synergies among various motion-relevant tasks. Text, the most familiar and well-understood modality for LLMs, is utilized as a bridge to establish connections between different motion tasks, facilitating mutual reinforcement. To our knowledge, M$^3$GPT is the first model capable of comprehending and generating motions based on multiple signals. Extensive experiments highlight M$^3$GPT's superior performance across various motion-relevant tasks and its powerful zero-shot generalization capabilities for extremely challenging tasks.
Authors: Junyoung Seo, Kazumi Fukuda, Takashi Shibuya, Takuya Narihira, Naoki Murata, Shoukang Hu, Chieh-Hsin Lai, Seungryong Kim, Yuki Mitsufuji
Abstract: Generating novel views from a single image remains a challenging task due to the complexity of 3D scenes and the limited diversity in the existing multi-view datasets to train a model on. Recent research combining large-scale text-to-image (T2I) models with monocular depth estimation (MDE) has shown promise in handling in-the-wild images. In these methods, an input view is geometrically warped to novel views with estimated depth maps, then the warped image is inpainted by T2I models. However, they struggle with noisy depth maps and loss of semantic details when warping an input view to novel viewpoints. In this paper, we propose a novel approach for single-shot novel view synthesis, a semantic-preserving generative warping framework that enables T2I generative models to learn where to warp and where to generate, through augmenting cross-view attention with self-attention. Our approach addresses the limitations of existing methods by conditioning the generative model on source view images and incorporating geometric warping signals. Qualitative and quantitative evaluations demonstrate that our model outperforms existing methods in both in-domain and out-of-domain scenarios. Project page is available at https://GenWarp-NVS.github.io/.
Authors: Khoa Vo, Thinh Phan, Kashu Yamazaki, Minh Tran, Ngan Le
Abstract: Current video-language models (VLMs) rely extensively on instance-level alignment between video and language modalities, which presents two major limitations: (1) visual reasoning disobeys the natural perception that humans do in first-person perspective, leading to a lack of reasoning interpretation; and (2) learning is limited in capturing inherent fine-grained relationships between two modalities. In this paper, we take an inspiration from human perception and explore a compositional approach for egocentric video representation. We introduce HENASY (Hierarchical ENtities ASsemblY), which includes a spatiotemporal token grouping mechanism to explicitly assemble dynamically evolving scene entities through time and model their relationship for video representation. By leveraging compositional structure understanding, HENASY possesses strong interpretability via visual grounding with free-form text queries. We further explore a suite of multi-grained contrastive losses to facilitate entity-centric understandings. This comprises three alignment types: video-narration, noun-entity, verb-entities alignments. Our method demonstrates strong interpretability in both quantitative and qualitative experiments; while maintaining competitive performances on five downstream tasks via zero-shot transfer or as video/text representation, including video/text retrieval, action recognition, multi-choice query, natural language query, and moments query.
Authors: Yue Lu, Shizhou Zhang, De Cheng, Yinghui Xing, Nannan Wang, Peng Wang, Yanning Zhang
Abstract: Existing prompt-tuning methods have demonstrated impressive performances in continual learning (CL), by selecting and updating relevant prompts in the vision-transformer models. On the contrary, this paper aims to learn each task by tuning the prompts in the direction orthogonal to the subspace spanned by previous tasks' features, so as to ensure no interference on tasks that have been learned to overcome catastrophic forgetting in CL. However, different from the orthogonal projection in the traditional CNN architecture, the prompt gradient orthogonal projection in the ViT architecture shows completely different and greater challenges, i.e., 1) the high-order and non-linear self-attention operation; 2) the drift of prompt distribution brought by the LayerNorm in the transformer block. Theoretically, we have finally deduced two consistency conditions to achieve the prompt gradient orthogonal projection, which provide a theoretical guarantee of eliminating interference on previously learned knowledge via the self-attention mechanism in visual prompt tuning. In practice, an effective null-space-based approximation solution has been proposed to implement the prompt gradient orthogonal projection. Extensive experimental results demonstrate the effectiveness of anti-forgetting on four class-incremental benchmarks with diverse pre-trained baseline models, and our approach achieves superior performances to state-of-the-art methods. Our code is available at https://github.com/zugexiaodui/VPTinNSforCL.
Authors: Zigeng Chen, Xinyin Ma, Gongfan Fang, Zhenxiong Tan, Xinchao Wang
Abstract: Diffusion models have garnered significant interest from the community for their great generative ability across various applications. However, their typical multi-step sequential-denoising nature gives rise to high cumulative latency, thereby precluding the possibilities of parallel computation. To address this, we introduce AsyncDiff, a universal and plug-and-play acceleration scheme that enables model parallelism across multiple devices. Our approach divides the cumbersome noise prediction model into multiple components, assigning each to a different device. To break the dependency chain between these components, it transforms the conventional sequential denoising into an asynchronous process by exploiting the high similarity between hidden states in consecutive diffusion steps. Consequently, each component is facilitated to compute in parallel on separate devices. The proposed strategy significantly reduces inference latency while minimally impacting the generative quality. Specifically, for the Stable Diffusion v2.1, AsyncDiff achieves a 2.7x speedup with negligible degradation and a 4.0x speedup with only a slight reduction of 0.38 in CLIP Score, on four NVIDIA A5000 GPUs. Our experiments also demonstrate that AsyncDiff can be readily applied to video diffusion models with encouraging performances. The code is available at https://github.com/czg1225/AsyncDiff.
Authors: Yihong Xu, \'Eloi Zablocki, Alexandre Boulch, Gilles Puy, Mickael Chen, Florent Bartoccioni, Nermin Samet, Oriane Sim\'eoni, Spyros Gidaris, Tuan-Hung Vu, Andrei Bursuc, Eduardo Valle, Renaud Marlet, Matthieu Cord
Abstract: Motion forecasting is crucial in autonomous driving systems to anticipate the future trajectories of surrounding agents such as pedestrians, vehicles, and traffic signals. In end-to-end forecasting, the model must jointly detect and track from sensor data (cameras or LiDARs) the past trajectories of the different elements of the scene and predict their future locations. We depart from the current trend of tackling this task via end-to-end training from perception to forecasting, and instead use a modular approach. We individually build and train detection, tracking and forecasting modules. We then only use consecutive finetuning steps to integrate the modules better and alleviate compounding errors. We conduct an in-depth study on the finetuning strategies and it reveals that our simple yet effective approach significantly improves performance on the end-to-end forecasting benchmark. Consequently, our solution ranks first in the Argoverse 2 End-to-end Forecasting Challenge, with 63.82 mAPf. We surpass forecasting results by +17.1 points over last year's winner and by +13.3 points over this year's runner-up. This remarkable performance in forecasting can be explained by our modular paradigm, which integrates finetuning strategies and significantly outperforms the end-to-end-trained counterparts. The code, model weights and results are made available https://github.com/valeoai/valeo4cast.
Authors: Mohamed Fadhlallah Guerri, Cosimo Distante, Paolo Spagnolo, Fares Bougourzi, Abdelmalik Taleb-Ahmed
Abstract: During the process of classifying Hyperspectral Image (HSI), every pixel sample is categorized under a land-cover type. CNN-based techniques for HSI classification have notably advanced the field by their adept feature representation capabilities. However, acquiring deep features remains a challenge for these CNN-based methods. In contrast, transformer models are adept at extracting high-level semantic features, offering a complementary strength. This paper's main contribution is the introduction of an HSI classification model that includes two convolutional blocks, a Gate-Shift-Fuse (GSF) block and a transformer block. This model leverages the strengths of CNNs in local feature extraction and transformers in long-range context modelling. The GSF block is designed to strengthen the extraction of local and global spatial-spectral features. An effective attention mechanism module is also proposed to enhance the extraction of information from HSI cubes. The proposed method is evaluated on four well-known datasets (the Indian Pines, Pavia University, WHU-WHU-Hi-LongKou and WHU-Hi-HanChuan), demonstrating that the proposed framework achieves superior results compared to other models.
Authors: Jiayi Liu, Qianyu Zhang, Xue Wan, Shengyang Zhang, Yaolin Tian, Haodong Han, Yutao Zhao, Baichuan Liu, Zeyuan Zhao, Xubo Luo
Abstract: With the complexity of lunar exploration missions, the moon needs to have a higher level of autonomy. Environmental perception and navigation algorithms are the foundation for lunar rovers to achieve autonomous exploration. The development and verification of algorithms require highly reliable data support. Most of the existing lunar datasets are targeted at a single task, lacking diverse scenes and high-precision ground truth labels. To address this issue, we propose a multi-task, multi-scene, and multi-label lunar benchmark dataset LuSNAR. This dataset can be used for comprehensive evaluation of autonomous perception and navigation systems, including high-resolution stereo image pairs, panoramic semantic labels, dense depth maps, LiDAR point clouds, and the position of rover. In order to provide richer scene data, we built 9 lunar simulation scenes based on Unreal Engine. Each scene is divided according to topographic relief and the density of objects. To verify the usability of the dataset, we evaluated and analyzed the algorithms of semantic segmentation, 3D reconstruction, and autonomous navigation. The experiment results prove that the dataset proposed in this paper can be used for ground verification of tasks such as autonomous environment perception and navigation, and provides a lunar benchmark dataset for testing the accessibility of algorithm metrics. We make LuSNAR publicly available at: https://github.com/zqyu9/LuSNAR-dataset.
Authors: Yibo Zhong, Yao Zhou
Abstract: Low-rank adaptation (LoRA) is a powerful parameter-efficient fine-tuning method that utilizes low-rank projectors $A$ and $B$ to learn weight updates $\Delta W$ for adaptation targets $W$. Previous research has shown that LoRA is essentially a gradient compressor, performing random projections on the gradient using a fixed projection matrix $A_0$. However, this setup restricts the overall weight update to be low-rank, which limits the adaptation performance. In this paper, we propose low-rank interconnected adaptation across layers (Lily). Specifically, we employ a hierarchical framework where low-dimensional projectors (LPs) retained for downward projection at a particular level, while globally-shared high-dimensional projector (HP) experts perform upward projection across all levels of layers. Lily uniquely connects each LP to all HP experts, therefore the gradient projections are no longer dominated by fixed projection matrices, but rather by selective combinations of all the projectors, thereby breaking the low-rank constraint of LoRA. Furthermore, Lily's cross-layer connections facilitate the capture of intricate information and dependencies across different layers, thereby enhancing the model's representational capabilities. Experiments across various modalities, architectures, and model sizes underscore Lily's great performance and efficiency. Code is available on github https://github.com/yibozhong/lily.
Authors: Yunling Zheng, Zeyi Xu, Fanghui Xue, Biao Yang, Jiancheng Lyu, Shuai Zhang, Yingyong Qi, Jack Xin
Abstract: We propose and demonstrate an alternating Fourier and image domain filtering approach for feature extraction as an efficient alternative to build a vision backbone without using the computationally intensive attention. The performance among the lightweight models reaches the state-of-the-art level on ImageNet-1K classification, and improves downstream tasks on object detection and segmentation consistently as well. Our approach also serves as a new tool to compress vision transformers (ViTs).
Authors: Ruizhe Wang, Chunliang Hua, Tomakayev Shingys, Mengyuan Niu, Qingxin Yang, Lizhong Gao, Yi Zheng, Junyan Yang, Qiao Wang
Abstract: The photorealistic reconstruction and rendering of architectural scenes have extensive applications in industries such as film, games, and transportation. It also plays an important role in urban planning, architectural design, and the city's promotion, especially in protecting historical and cultural relics. The 3D Gaussian Splatting, due to better performance over NeRF, has become a mainstream technology in 3D reconstruction. Its only input is a set of images but it relies heavily on geometric parameters computed by the SfM process. At the same time, there is an existing abundance of raw 3D models, that could inform the structural perception of certain buildings but cannot be applied. In this paper, we propose a straightforward method to harness these raw 3D models to guide 3D Gaussians in capturing the basic shape of the building and improve the visual quality of textures and details when photos are captured non-systematically. This exploration opens up new possibilities for improving the effectiveness of 3D reconstruction techniques in the field of architectural design.
Authors: Pieter M. Blok, Federico Magistri, Cyrill Stachniss, Haozhou Wang, James Burridge, Wei Guo
Abstract: Potato yield is an important metric for farmers to further optimize their cultivation practices. Potato yield can be estimated on a harvester using an RGB-D camera that can estimate the three-dimensional (3D) volume of individual potato tubers. A challenge, however, is that the 3D shape derived from RGB-D images is only partially completed, underestimating the actual volume. To address this issue, we developed a 3D shape completion network, called CoRe++, which can complete the 3D shape from RGB-D images. CoRe++ is a deep learning network that consists of a convolutional encoder and a decoder. The encoder compresses RGB-D images into latent vectors that are used by the decoder to complete the 3D shape using the deep signed distance field network (DeepSDF). To evaluate our CoRe++ network, we collected partial and complete 3D point clouds of 339 potato tubers on an operational harvester in Japan. On the 1425 RGB-D images in the test set (representing 51 unique potato tubers), our network achieved a completion accuracy of 2.8 mm on average. For volumetric estimation, the root mean squared error (RMSE) was 22.6 ml, and this was better than the RMSE of the linear regression (31.1 ml) and the base model (36.9 ml). We found that the RMSE can be further reduced to 18.2 ml when performing the 3D shape completion in the center of the RGB-D image. With an average 3D shape completion time of 10 milliseconds per tuber, we can conclude that CoRe++ is both fast and accurate enough to be implemented on an operational harvester for high-throughput potato yield estimation. Our method can also be applied to other tuber, fruit and vegetable crops, thereby enabling versatile, accurate and real-time yield monitoring in precision agriculture. Our code, network weights and dataset are publicly available at https://github.com/UTokyo-FieldPhenomics-Lab/corepp.git.
URLs: https://github.com/UTokyo-FieldPhenomics-Lab/corepp.git.
Authors: Jie Gui, Chengze Jiang, Minjing Dong, Kun Tong, Xinli Shi, Yuan Yan Tang, Dacheng Tao
Abstract: While adversarial training is an effective defense method against adversarial attacks, it notably increases the training cost. To this end, fast adversarial training (FAT) is presented for efficient training and has become a hot research topic. However, FAT suffers from catastrophic overfitting, which leads to a performance drop compared with multi-step adversarial training. However, the cause of catastrophic overfitting remains unclear and lacks exploration. In this paper, we present an example taxonomy in FAT, which identifies that catastrophic overfitting is caused by the imbalance between the inner and outer optimization in FAT. Furthermore, we investigated the impact of varying degrees of training loss, revealing a correlation between training loss and catastrophic overfitting. Based on these observations, we redesign the loss function in FAT with the proposed dynamic label relaxation to concentrate the loss range and reduce the impact of misclassified examples. Meanwhile, we introduce batch momentum initialization to enhance the diversity to prevent catastrophic overfitting in an efficient manner. Furthermore, we also propose Catastrophic Overfitting aware Loss Adaptation (COLA), which employs a separate training strategy for examples based on their loss degree. Our proposed method, named example taxonomy aware FAT (ETA), establishes an improved paradigm for FAT. Experiment results demonstrate our ETA achieves state-of-the-art performance. Comprehensive experiments on four standard datasets demonstrate the competitiveness of our proposed method.
Authors: Lukas Meyer, Andreas Gilson, Ute Schmid, Marc Stamminger
Abstract: We introduce FruitNeRF, a unified novel fruit counting framework that leverages state-of-the-art view synthesis methods to count any fruit type directly in 3D. Our framework takes an unordered set of posed images captured by a monocular camera and segments fruit in each image. To make our system independent of the fruit type, we employ a foundation model that generates binary segmentation masks for any fruit. Utilizing both modalities, RGB and semantic, we train a semantic neural radiance field. Through uniform volume sampling of the implicit Fruit Field, we obtain fruit-only point clouds. By applying cascaded clustering on the extracted point cloud, our approach achieves precise fruit count.The use of neural radiance fields provides significant advantages over conventional methods such as object tracking or optical flow, as the counting itself is lifted into 3D. Our method prevents double counting fruit and avoids counting irrelevant fruit.We evaluate our methodology using both real-world and synthetic datasets. The real-world dataset consists of three apple trees with manually counted ground truths, a benchmark apple dataset with one row and ground truth fruit location, while the synthetic dataset comprises various fruit types including apple, plum, lemon, pear, peach, and mango.Additionally, we assess the performance of fruit counting using the foundation model compared to a U-Net.
Authors: Ziyu Tang, Weicai Ye, Yifan Wang, Di Huang, Hujun Bao, Tong He, Guofeng Zhang
Abstract: Neural implicit reconstruction via volume rendering has demonstrated its effectiveness in recovering dense 3D surfaces. However, it is non-trivial to simultaneously recover meticulous geometry and preserve smoothness across regions with differing characteristics. To address this issue, previous methods typically employ geometric priors, which are often constrained by the performance of the prior models. In this paper, we propose ND-SDF, which learns a Normal Deflection field to represent the angular deviation between the scene normal and the prior normal. Unlike previous methods that uniformly apply geometric priors on all samples, introducing significant bias in accuracy, our proposed normal deflection field dynamically learns and adapts the utilization of samples based on their specific characteristics, thereby improving both the accuracy and effectiveness of the model. Our method not only obtains smooth weakly textured regions such as walls and floors but also preserves the geometric details of complex structures. In addition, we introduce a novel ray sampling strategy based on the deflection angle to facilitate the unbiased rendering process, which significantly improves the quality and accuracy of intricate surfaces, especially on thin structures. Consistent improvements on various challenging datasets demonstrate the superiority of our method.
Authors: Gonglin Chen, Jinsen Wu, Haiwei Chen, Wenbin Teng, Zhiyuan Gao, Andrew Feng, Rongjun Qin, Yajie Zhao
Abstract: Establishing consistent and dense correspondences across multiple images is crucial for Structure from Motion (SfM) systems. Significant view changes, such as air-to-ground with very sparse view overlap, pose an even greater challenge to the correspondence solvers. We present a novel optimization-based approach that significantly enhances existing feature matching methods by introducing geometry cues in addition to color cues. This helps fill gaps when there is less overlap in large-scale scenarios. Our method formulates geometric verification as an optimization problem, guiding feature matching within detector-free methods and using sparse correspondences from detector-based methods as anchor points. By enforcing geometric constraints via the Sampson Distance, our approach ensures that the denser correspondences from detector-free methods are geometrically consistent and more accurate. This hybrid strategy significantly improves correspondence density and accuracy, mitigates multi-view inconsistencies, and leads to notable advancements in camera pose accuracy and point cloud density. It outperforms state-of-the-art feature matching methods on benchmark datasets and enables feature matching in challenging extreme large-scale settings.
Authors: Jiahua Dong, Yue Zhang, Qiuli Wang, Ruofeng Tong, Shihong Ying, Shaolin Gong, Xuanpu Zhang, Lanfen Lin, Yen-Wei Chen, S. Kevin Zhou
Abstract: Medical image segmentation is crucial in the field of medical imaging, aiding in disease diagnosis and surgical planning. Most established segmentation methods rely on supervised deep learning, in which clean and precise labels are essential for supervision and significantly impact the performance of models. However, manually delineated labels often contain noise, such as missing labels and inaccurate boundary delineation, which can hinder networks from correctly modeling target characteristics. In this paper, we propose a deep self-cleansing segmentation framework that can preserve clean labels while cleansing noisy ones in the training phase. To achieve this, we devise a gaussian mixture model-based label filtering module that distinguishes noisy labels from clean labels. Additionally, we develop a label cleansing module to generate pseudo low-noise labels for identified noisy samples. The preserved clean labels and pseudo-labels are then used jointly to supervise the network. Validated on a clinical liver tumor dataset and a public cardiac diagnosis dataset, our method can effectively suppress the interference from noisy labels and achieve prominent segmentation performance.
Authors: Yang Li, Quan Yuan, Guiyang Luo, Xiaoyuan Fu, Xuanhan Zhu, Yujia Yang, Rui Pan, Jinglin Li
Abstract: By sharing complementary perceptual information, multi-agent collaborative perception fosters a deeper understanding of the environment. Recent studies on collaborative perception mostly utilize CNNs or Transformers to learn feature representation and fusion in the spatial dimension, which struggle to handle long-range spatial-temporal features under limited computing and communication resources. Holistically modeling the dependencies over extensive spatial areas and extended temporal frames is crucial to enhancing feature quality. To this end, we propose a resource efficient cross-agent spatial-temporal collaborative state space model (SSM), named CollaMamba. Initially, we construct a foundational backbone network based on spatial SSM. This backbone adeptly captures positional causal dependencies from both single-agent and cross-agent views, yielding compact and comprehensive intermediate features while maintaining linear complexity. Furthermore, we devise a history-aware feature boosting module based on temporal SSM, extracting contextual cues from extended historical frames to refine vague features while preserving low overhead. Extensive experiments across several datasets demonstrate that CollaMamba outperforms state-of-the-art methods, achieving higher model accuracy while reducing computational and communication overhead by up to 71.9% and 1/64, respectively. This work pioneers the exploration of the Mamba's potential in collaborative perception. The source code will be made available.
Authors: Pei Liu, Luping Ji, Jiaxiang Gou, Bo Fu, Mao Ye
Abstract: Histopathology Whole-Slide Images (WSIs) provide an important tool to assess cancer prognosis in computational pathology (CPATH). While existing survival analysis (SA) approaches have made exciting progress, they are generally limited to adopting highly-expressive architectures and only coarse-grained patient-level labels to learn prognostic visual representations from gigapixel WSIs. Such learning paradigm suffers from important performance bottlenecks, when facing present scarce training data and standard multi-instance learning (MIL) framework in CPATH. To overcome it, this paper, for the first time, proposes a new Vision-Language-based SA (VLSA) paradigm. Concretely, (1) VLSA is driven by pathology VL foundation models. It no longer relies on high-capability networks and shows the advantage of data efficiency. (2) In vision-end, VLSA encodes prognostic language prior and then employs it as auxiliary signals to guide the aggregating of prognostic visual features at instance level, thereby compensating for the weak supervision in MIL. Moreover, given the characteristics of SA, we propose i) ordinal survival prompt learning to transform continuous survival labels into textual prompts; and ii) ordinal incidence function as prediction target to make SA compatible with VL-based prediction. Notably, VLSA's predictions can be interpreted intuitively by our Shapley values-based method. The extensive experiments on five datasets confirm the effectiveness of our scheme. Our VLSA could pave a new way for SA in CPATH by offering weakly-supervised MIL an effective means to learn valuable prognostic clues from gigapixel WSIs. Our source code is available at https://github.com/liupei101/VLSA.
Authors: Lulin Zhang, Ewelina Rupnik, Tri Dung Nguyen, St\'ephane Jacquemoud, Yann Klinger
Abstract: Neural radiance fields (NeRF) have gained prominence as a machine learning technique for representing 3D scenes and estimating the bidirectional reflectance distribution function (BRDF) from multiple images. However, most existing research has focused on close-range imagery, typically modeling scene surfaces with simplified Microfacet BRDF models, which are often inadequate for representing complex Earth surfaces. Furthermore, NeRF approaches generally require large sets of simultaneously captured images for high-quality surface depth reconstruction - a condition rarely met in satellite imaging. To overcome these challenges, we introduce BRDF-NeRF, which incorporates the physically-based semi-empirical Rahman-Pinty-Verstraete (RPV) BRDF model, known to better capture the reflectance properties of natural surfaces. Additionally, we propose guided volumetric sampling and depth supervision to enable radiance field modeling with a minimal number of views. Our method is evaluated on two satellite datasets: (1) Djibouti, captured at varying viewing angles within a single epoch with a fixed Sun position, and (2) Lanzhou, captured across multiple epochs with different Sun positions and viewing angles. Using only three to four satellite images for training, BRDF-NeRF successfully synthesizes novel views from unseen angles and generates high-quality digital surface models (DSMs).
Authors: Avi Deb Raha, Apurba Adhikary, Mrityunjoy Gain, Yu Qiao, Choong Seon Hong
Abstract: In this study, we explore the efficacy of advanced pre-trained architectures, such as Vision Transformers (ViT), ConvNeXt, and Swin Transformers in enhancing Federated Domain Generalization. These architectures capture global contextual features and model long-range dependencies, making them promising candidates for improving cross-domain generalization. We conduct a broad study with in-depth analysis and systematically evaluate different variants of these architectures, using extensive pre-training datasets such as ImageNet-1K, ImageNet-21K, JFT-300M, and ImageNet-22K. Additionally, we compare self-supervised and supervised pre-training strategies to assess their impact on FDG performance. Our findings suggest that self-supervised techniques, which focus on reconstructing masked image patches, can better capture the intrinsic structure of images, thereby outperforming their supervised counterparts. Comprehensive evaluations on the Office-Home and PACS datasets demonstrate that adopting advanced architectures pre-trained on larger datasets establishes new benchmarks, achieving average accuracies of 84.46\% and 92.55\%, respectively. Additionally, we observe that certain variants of these advanced models, despite having fewer parameters, outperform larger ResNet models. This highlights the critical role of utilizing sophisticated architectures and diverse pre-training strategies to enhance FDG performance, especially in scenarios with limited computational resources where model efficiency is crucial. Our results indicate that federated learning systems can become more adaptable and efficient by leveraging these advanced methods, offering valuable insights for future research in FDG.
Authors: Yalong Jiang, Liquan Mao
Abstract: Video anomaly detection is a subject of great interest across industrial and academic domains due to its crucial role in computer vision applications. However, the inherent unpredictability of anomalies and the scarcity of anomaly samples present significant challenges for unsupervised learning methods. To overcome the limitations of unsupervised learning, which stem from a lack of comprehensive prior knowledge about anomalies, we propose VLAVAD (Video-Language Models Assisted Anomaly Detection). Our method employs a cross-modal pre-trained model that leverages the inferential capabilities of large language models (LLMs) in conjunction with a Selective-Prompt Adapter (SPA) for selecting semantic space. Additionally, we introduce a Sequence State Space Module (S3M) that detects temporal inconsistencies in semantic features. By mapping high-dimensional visual features to low-dimensional semantic ones, our method significantly enhance the interpretability of unsupervised anomaly detection. Our proposed approach effectively tackles the challenge of detecting elusive anomalies that are hard to discern over periods, achieving SOTA on the challenging ShanghaiTech dataset.
Authors: Tomasz Stanczyk, Francois Bremond
Abstract: Multi-object tracking (MOT) involves identifying and consistently tracking objects across video sequences. Traditional tracking-by-detection methods, while effective, often require extensive tuning and lack generalizability. On the other hand, segmentation mask-based methods are more generic but struggle with tracking management, making them unsuitable for MOT. We propose a novel approach, McByte, which incorporates a temporally propagated segmentation mask as a strong association cue within a tracking-by-detection framework. By combining bounding box and mask information, McByte enhances robustness and generalizability without per-sequence tuning. Evaluated on four benchmark datasets - DanceTrack, MOT17, SoccerNet-tracking 2022, and KITTI-tracking - McByte demonstrates performance gain in all cases examined. At the same time, it outperforms existing mask-based methods. Implementation code will be provided upon acceptance.
Authors: Peizhi Yan, Rabab Ward, Qiang Tang, Shan Du
Abstract: Recent advancements in 3D Gaussian Splatting (3DGS) have unlocked significant potential for modeling 3D head avatars, providing greater flexibility than mesh-based methods and more efficient rendering compared to NeRF-based approaches. Despite these advancements, the creation of controllable 3DGS-based head avatars remains time-intensive, often requiring tens of minutes to hours. To expedite this process, we here introduce the ``Gaussian D\'ej\`a-vu" framework, which first obtains a generalized model of the head avatar and then personalizes the result. The generalized model is trained on large 2D (synthetic and real) image datasets. This model provides a well-initialized 3D Gaussian head that is further refined using a monocular video to achieve the personalized head avatar. For personalizing, we propose learnable expression-aware rectification blendmaps to correct the initial 3D Gaussians, ensuring rapid convergence without the reliance on neural networks. Experiments demonstrate that the proposed method meets its objectives. It outperforms state-of-the-art 3D Gaussian head avatars in terms of photorealistic quality as well as reduces training time consumption to at least a quarter of the existing methods, producing the avatar in minutes.
Authors: Sunghyun Ahn, Youngwan Jo, Kijung Lee, Sanghyun Park
Abstract: Video anomaly detection (VAD) is a crucial task in video analysis and surveillance within computer vision. Currently, VAD is gaining attention with memory techniques that store the features of normal frames. The stored features are utilized for frame reconstruction, identifying an abnormality when a significant difference exists between the reconstructed and input frames. However, this approach faces several challenges due to the simultaneous optimization required for both the memory and encoder-decoder model. These challenges include increased optimization difficulty, complexity of implementation, and performance variability depending on the memory size. To address these challenges,we propose an effective memory method for VAD, called VideoPatchCore. Inspired by PatchCore, our approach introduces a structure that prioritizes memory optimization and configures three types of memory tailored to the characteristics of video data. This method effectively addresses the limitations of existing memory-based methods, achieving good performance comparable to state-of-the-art methods. Furthermore, our method requires no training and is straightforward to implement, making VAD tasks more accessible. Our code is available online at github.com/SkiddieAhn/Paper-VideoPatchCore.
Authors: Jiacheng Zhang, Yang Jiao, Shaoxiang Chen, Jingjing Chen, Yu-Gang Jiang
Abstract: Recently, Multimodal Large Language Models (MLLMs) have sparked great research interests owing to their exceptional content-reasoning and instruction-following capabilities. To effectively instruct an MLLM, in addition to conventional language expressions, the practice of referring to objects by painting with brushes on images has emerged as a prevalent tool (referred to as "referring visual prompts") due to its efficacy in aligning the user's intention with specific image regions. To accommodate the most common referring visual prompts, namely points, boxes, and masks, existing approaches initially utilize specialized feature encoding modules to capture the semantics of the highlighted areas indicated by these prompts. Subsequently, these encoded region features are adapted to MLLMs through fine-tuning on a meticulously curated multimodal instruction dataset. However, such designs suffer from redundancy in architecture. Moreover, they face challenges in effectively generalizing when encountering a diverse range of arbitrary referring visual prompts in real-life scenarios. To address the above issues, we propose EAGLE, a novel MLLM that empowers comprehension of arbitrary referring visual prompts with less training efforts than existing approaches. Specifically, our EAGLE maintains the innate format of the referring visual prompts as colored patches rendered on the given image for conducting the instruction tuning. Our approach embeds referring visual prompts as spatial concepts conveying specific spatial areas comprehensible to the MLLM, with the semantic comprehension of these regions originating from the MLLM itself. Besides, we also propose a Geometry-Agnostic Learning paradigm (GAL) to further disentangle the MLLM's region-level comprehension with the specific formats of referring visual prompts. Extensive experiments are conducted to prove the effectiveness of our proposed method.
Authors: Jacob Fein-Ashley, Ethan Feng, Minh Pham
Abstract: Data representation in non-Euclidean spaces has proven effective for capturing hierarchical and complex relationships in real-world datasets. Hyperbolic spaces, in particular, provide efficient embeddings for hierarchical structures. This paper introduces the Hyperbolic Vision Transformer (HVT), a novel extension of the Vision Transformer (ViT) that integrates hyperbolic geometry. While traditional ViTs operate in Euclidean space, our method enhances the self-attention mechanism by leveraging hyperbolic distance and M\"obius transformations. This enables more effective modeling of hierarchical and relational dependencies in image data. We present rigorous mathematical formulations, showing how hyperbolic geometry can be incorporated into attention layers, feed-forward networks, and optimization. We offer improved performance for image classification using the ImageNet dataset.
Authors: Kunchi Li, Jun Wan, Shan Yu
Abstract: Rehearsal approaches in class incremental learning (CIL) suffer from decision boundary overfitting to new classes, which is mainly caused by two factors: insufficiency of old classes data for knowledge distillation and imbalanced data learning between the learned and new classes because of the limited storage memory. In this work, we present a simple but effective approach to tackle these two factors. First, we employ a re-sampling strategy and Mixup K}nowledge D}istillation (Re-MKD) to improve the performances of KD, which would greatly alleviate the overfitting problem. Specifically, we combine mixup and re-sampling strategies to synthesize adequate data used in KD training that are more consistent with the latent distribution between the learned and new classes. Second, we propose a novel incremental influence balance (IIB) method for CIL to tackle the classification of imbalanced data by extending the influence balance method into the CIL setting, which re-weights samples by their influences to create a proper decision boundary. With these two improvements, we present the effective decision boundary learning algorithm (EDBL) which improves the performance of KD and deals with the imbalanced data learning simultaneously. Experiments show that the proposed EDBL achieves state-of-the-art performances on several CIL benchmarks.
Authors: Ana-Maria Marcu, Long Chen, Jan H\"unermann, Alice Karnsund, Benoit Hanotte, Prajwal Chidananda, Saurabh Nair, Vijay Badrinarayanan, Alex Kendall, Jamie Shotton, Elahe Arani, Oleg Sinavski
Abstract: We introduce LingoQA, a novel dataset and benchmark for visual question answering in autonomous driving. The dataset contains 28K unique short video scenarios, and 419K annotations. Evaluating state-of-the-art vision-language models on our benchmark shows that their performance is below human capabilities, with GPT-4V responding truthfully to 59.6% of the questions compared to 96.6% for humans. For evaluation, we propose a truthfulness classifier, called Lingo-Judge, that achieves a 0.95 Spearman correlation coefficient to human evaluations, surpassing existing techniques like METEOR, BLEU, CIDEr, and GPT-4. We establish a baseline vision-language model and run extensive ablation studies to understand its performance. We release our dataset and benchmark as an evaluation platform for vision-language models in autonomous driving.
Authors: Kenta Tsukahara, Kanji Tanaka
Abstract: A typical assumption in state-of-the-art self-localization models is that an annotated training dataset is available for the target workspace. However, this is not necessarily true when a robot travels around the general open world. This work introduces a novel training scheme for open-world distributed robot systems. In our scheme, a robot (``student") can ask the other robots it meets at unfamiliar places (``teachers") for guidance. Specifically, a pseudo-training dataset is reconstructed from the teacher model and then used for continual learning of the student model under domain, class, and vocabulary incremental setup. Unlike typical knowledge transfer schemes, our scheme introduces only minimal assumptions on the teacher model, so that it can handle various types of open-set teachers, including those uncooperative, untrainable (e.g., image retrieval engines), or black-box teachers (i.e., data privacy). In this paper, we investigate a ranking function as an instance of such generic models, using a challenging data-free recursive distillation scenario, where a student once trained can recursively join the next-generation open teacher set.
Authors: Huan-Ke Hsu, I-Chyn Wey, T. Hui Teo
Abstract: Generative Artificial Intelligence (AI) has become incredibly popular in recent years, and the significance of traditional accelerators in dealing with large-scale parameters is urgent. With the diffusion model's parallel structure, the hardware design challenge has skyrocketed because of the multiple layers operating simultaneously. Convolution Neural Network (CNN) accelerators have been designed and developed rapidly, especially for high-speed inference. Often, CNN models with parallel structures are deployed. In these CNN accelerators, many Processing Elements (PE) are required to perform parallel computations, mainly the multiply and accumulation (MAC) operation, resulting in high power consumption and a large silicon area. In this work, a Server Flow Multi-Mode CNN Unit (SF-MMCN) is proposed to reduce the number of PE while improving the operation efficiency of the CNN accelerator. The pipelining technique is introduced into Server Flow to process parallel computations. The proposed SF-MMCN is implemented with TSMC 90-nm CMOS technology. It is evaluated with VGG-16, ResNet-18, and U-net. The evaluation results show that the proposed SF-MMCN can reduce the power consumption by 92%, and the silicon area by 70%, while improving the efficiency of operation by nearly 81 times. A new FoM, area efficiency (GOPs/mm^2) is also introduced to evaluate the performance of the accelerator in terms of the ratio throughput (GOPs) and silicon area (mm^2). In this FoM, SF-MMCN improves area efficiency by 18 times (18.42).
Authors: Haosong Peng, Wei Feng, Hao Li, Yufeng Zhan, Ren Jin, Yuanqing Xia
Abstract: The advent of edge computing has made real-time intelligent video analytics feasible. Previous works, based on traditional model architecture (e.g., CNN, RNN, etc.), employ various strategies to filter out non-region-of-interest content to minimize bandwidth and computation consumption but show inferior performance in adverse environments. Recently, visual foundation models based on transformers have shown great performance in adverse environments due to their amazing generalization capability. However, they require a large amount of computation power, which limits their applications in real-time intelligent video analytics. In this paper, we find visual foundation models like Vision Transformer (ViT) also have a dedicated acceleration mechanism for video analytics. To this end, we introduce Arena, an end-to-end edge-assisted video inference acceleration system based on ViT. We leverage the capability of ViT that can be accelerated through token pruning by only offloading and feeding Patches-of-Interest to the downstream models. Additionally, we design an adaptive keyframe inference switching algorithm tailored to different videos, capable of adapting to the current video content to jointly optimize accuracy and bandwidth. Through extensive experiments, our findings reveal that Arena can boost inference speeds by up to 1.58\(\times\) and 1.82\(\times\) on average while consuming only 47\% and 31\% of the bandwidth, respectively, all with high inference accuracy.
Authors: Kaixin Li, Yuchen Tian, Qisheng Hu, Ziyang Luo, Zhiyong Huang, Jing Ma
Abstract: Programming often involves converting detailed and complex specifications into code, a process during which developers typically utilize visual aids to more effectively convey concepts. While recent developments in Large Multimodal Models have demonstrated remarkable abilities in visual reasoning and mathematical tasks, there is little work on investigating whether these models can effectively interpret visual elements for code generation. To this end, we present MMCode, the first multi-modal coding dataset for evaluating algorithmic problem-solving skills in visually rich contexts. MMCode contains 3,548 questions and 6,620 images collected from real-world programming challenges harvested from 10 code competition websites, presenting significant challenges due to the extreme demand for reasoning abilities. Our experiment results show that current state-of-the-art models struggle to solve these problems. The results highlight the lack of powerful vision-code models, and we hope MMCode can serve as an inspiration for future works in this domain. The data and code are publicly available at https://github.com/likaixin2000/MMCode.
Authors: Michelle Espranita Liman, Daniel Rueckert, Florian J. Fintelmann, Philip M\"uller
Abstract: Field-of-view (FOV) recovery of truncated chest CT scans is crucial for accurate body composition analysis, which involves quantifying skeletal muscle and subcutaneous adipose tissue (SAT) on CT slices. This, in turn, enables disease prognostication. Here, we present a method for recovering truncated CT slices using generative image outpainting. We train a diffusion model and apply it to truncated CT slices generated by simulating a small FOV. Our model reliably recovers the truncated anatomy and outperforms the previous state-of-the-art despite being trained on 87% less data.
Authors: Tong Zhang, Yingdong Hu, Jiacheng You, Yang Gao
Abstract: Given the high cost of collecting robotic data in the real world, sample efficiency is a consistently compelling pursuit in robotics. In this paper, we introduce SGRv2, an imitation learning framework that enhances sample efficiency through improved visual and action representations. Central to the design of SGRv2 is the incorporation of a critical inductive bias-action locality, which posits that robot's actions are predominantly influenced by the target object and its interactions with the local environment. Extensive experiments in both simulated and real-world settings demonstrate that action locality is essential for boosting sample efficiency. SGRv2 excels in RLBench tasks with keyframe control using merely 5 demonstrations and surpasses the RVT baseline in 23 of 26 tasks. Furthermore, when evaluated on ManiSkill2 and MimicGen using dense control, SGRv2's success rate is 2.54 times that of SGR. In real-world environments, with only eight demonstrations, SGRv2 can perform a variety of tasks at a markedly higher success rate compared to baseline models. Project website: http://sgrv2-robot.github.io
Authors: Salil B Patel, Vicky Goh, James F FitzGerald, Chrystalina A Antoniades
Abstract: Parkinson's Disease (PD) diagnosis remains challenging. This study applies Convolutional Kolmogorov-Arnold Networks (ConvKANs), integrating learnable spline-based activation functions into convolutional layers, for PD classification using structural MRI. The first 3D implementation of ConvKANs for medical imaging is presented, comparing their performance to Convolutional Neural Networks (CNNs) and Graph Convolutional Networks (GCNs) across three open-source datasets. Isolated analyses assessed performance within individual datasets, using cross-validation techniques. Holdout analyses evaluated cross-dataset generalizability by training models on two datasets and testing on the third, mirroring real-world clinical scenarios. In isolated analyses, 2D ConvKANs achieved the highest AUC of 0.99 (95% CI: 0.98-0.99) on the PPMI dataset, outperforming 2D CNNs (AUC: 0.97, p = 0.0092). 3D models showed promise, with 3D CNN and 3D ConvKAN reaching an AUC of 0.85 on PPMI. In holdout analyses, 3D ConvKAN demonstrated superior generalization, achieving an AUC of 0.85 on early-stage PD data. GCNs underperformed in 2D but improved in 3D implementations. These findings highlight ConvKANs' potential for PD detection, emphasize the importance of 3D analysis in capturing subtle brain changes, and underscore cross-dataset generalization challenges. This study advances AI-assisted PD diagnosis using structural MRI and emphasizes the need for larger-scale validation.
Authors: Gilles Van De Vyver, Svein-Erik M{\aa}s{\o}y, H{\aa}vard Dalen, Bj{\o}rnar Leangen Grenne, Espen Holte, Sindre Hellum Olaisen, John Nyberg, Andreas {\O}stvik, Lasse L{\o}vstakken, Erik Smistad
Abstract: Automatic estimation of cardiac ultrasound image quality can be beneficial for guiding operators and ensuring the accuracy of clinical measurements. Previous work often fails to distinguish the view correctness of the echocardiogram from the image quality. Additionally, previous studies only provide a global image quality value, which limits their practical utility. In this work, we developed and compared three methods to estimate image quality: 1) classic pixel-based metrics like the generalized contrast-to-noise ratio (gCNR) on myocardial segments as region of interest and left ventricle lumen as background, obtained using a U-Net segmentation 2) local image coherence derived from a U-Net model that predicts coherence from B-Mode images 3) a deep convolutional network that predicts the quality of each region directly in an end-to-end fashion. We evaluate each method against manual regional image quality annotations by three experienced cardiologists. The results indicate poor performance of the gCNR metric, with Spearman correlation to the annotations of rho = 0.24. The end-to-end learning model obtains the best result, rho = 0.69, comparable to the inter-observer correlation, rho = 0.63. Finally, the coherence-based method, with rho = 0.58, outperformed the classical metrics and is more generic than the end-to-end approach. The image quality prediction tool is available as an open source Python library at https://github.com/GillesVanDeVyver/arqee.
Authors: Yuhao Zhu, Ethan Chen, Colin Hascup, Yukang Yan, Gaurav Sharma
Abstract: We propose an assistive technology that helps individuals with Color Vision Deficiencies (CVD) to recognize/name colors. A dichromat's color perception is a reduced two-dimensional (2D) subset of a normal trichromat's three dimensional color (3D) perception, leading to confusion when visual stimuli that appear identical to the dichromat are referred to by different color names. Using our proposed system, CVD individuals can interactively induce distinct perceptual changes to originally confusing colors via a computational color space transformation. By combining their original 2D precepts for colors with the discriminative changes, a three dimensional color space is reconstructed, where the dichromat can learn to resolve color name confusions and accurately recognize colors. Our system is implemented as an Augmented Reality (AR) interface on smartphones, where users interactively control the rotation through swipe gestures and observe the induced color shifts in the camera view or in a displayed image. Through psychophysical experiments and a longitudinal user study, we demonstrate that such rotational color shifts have discriminative power (initially confusing colors become distinct under rotation) and exhibit structured perceptual shifts dichromats can learn with modest training. The AR App is also evaluated in two real-world scenarios (building with lego blocks and interpreting artistic works); users all report positive experience in using the App to recognize object colors that they otherwise could not.
Authors: Ardhendu Sekhar, Vrinda Goel, Garima Jain, Abhijeet Patil, Ravi Kant Gupta, Tripti Bameta, Swapnil Rane, Amit Sethi
Abstract: The current standard for detecting human epidermal growth factor receptor 2 (HER2) status in breast cancer patients relies on HER2 amplification, identified through fluorescence in situ hybridization (FISH) or immunohistochemistry (IHC). However, hematoxylin and eosin (H\&E) tumor stains are more widely available, and accurately predicting HER2 status using H\&E could reduce costs and expedite treatment selection. Deep Learning algorithms for H&E have shown effectiveness in predicting various cancer features and clinical outcomes, including moderate success in HER2 status prediction. In this work, we employed a customized weak supervision classification technique combined with MoCo-v2 contrastive learning to predict HER2 status. We trained our pipeline on 182 publicly available H&E Whole Slide Images (WSIs) from The Cancer Genome Atlas (TCGA), for which annotations by the pathology team at Yale School of Medicine are publicly available. Our pipeline achieved an Area Under the Curve (AUC) of 0.85 across four different test folds. Additionally, we tested our model on 44 H&E slides from the TCGA-BRCA dataset, which had an HER2 score of 2+ and included corresponding HER2 status and FISH test results. These cases are considered equivocal for IHC, requiring an expensive FISH test on their IHC slides for disambiguation. Our pipeline demonstrated an AUC of 0.81 on these challenging H&E slides. Reducing the need for FISH test can have significant implications in cancer treatment equity for underserved populations.
Authors: Anqi Zhang, Wilson S. Geisler
Abstract: Visual search is a fundamental natural task for humans and other animals. We investigated the decision processes humans use when searching briefly presented displays having well-separated potential target-object locations. Performance was compared with the Bayesian-optimal decision process under the assumption that the information from the different potential target locations is statistically independent. Surprisingly, humans performed slightly better than optimal, despite humans' substantial loss of sensitivity in the fovea, and the implausibility of the human brain replicating the optimal computations. We show that three factors can quantitatively explain these seemingly paradoxical results. Most importantly, simple and fixed heuristic decision rules reach near optimal search performance. Secondly, foveal neglect primarily affects only the central potential target location. Finally, spatially correlated neural noise causes search performance to exceed that predicted for independent noise. These findings have far-reaching implications for understanding visual search tasks and other identification tasks in humans and other animals.
Authors: Hong-fu Chou, Vu Nguyen Ha, Prabhu Thiruvasagam, Thanh-Dung Le, Geoffrey Eappen, Ti Ti Nguyen, Luis M. Garces-Socarras, Jorge L. Gonzalez-Rios, Juan Carlos Merlano-Duncan, Symeon Chatzinotas
Abstract: Earth Observation (EO) systems play a crucial role in achieving Sustainable Development Goals by collecting and analyzing vital global data through satellite networks. These systems are essential for tasks like mapping, disaster monitoring, and resource management, but they face challenges in processing and transmitting large volumes of EO data, especially in specialized fields such as agriculture and real-time disaster response. Domain-adapted Large Language Models (LLMs) provide a promising solution by facilitating data fusion between extensive EO data and semantic EO data. By improving integration and interpretation of diverse datasets, LLMs address the challenges of processing specialized information in agriculture and disaster response applications. This fusion enhances the accuracy and relevance of transmitted data. This paper presents a framework for semantic communication in EO satellite networks, aimed at improving data transmission efficiency and overall system performance through cognitive processing techniques. The proposed system employs Discrete-Task-Oriented Source-Channel Coding (DT-JSCC) and Semantic Data Augmentation (SA) to focus on relevant information while minimizing communication overhead. By integrating cognitive semantic processing and inter-satellite links, the framework enhances the analysis and transmission of multispectral satellite imagery, improving object detection, pattern recognition, and real-time decision-making. The introduction of Cognitive Semantic Augmentation (CSA) allows satellites to process and transmit semantic information, boosting adaptability to changing environments and application needs. This end-to-end architecture is tailored for next-generation satellite networks, such as those supporting 6G, and demonstrates significant improvements in efficiency and accuracy.
Authors: Wanchen Zhao
Abstract: Image-to-Image Translation is a vital area of computer vision that focuses on transforming images from one visual domain to another while preserving their core content and structure. However, this field faces two major challenges: first, the data from the two domains are often unpaired, making it difficult to train generative adversarial networks effectively; second, existing methods tend to produce artifacts or hallucinations during image generation, leading to a decline in image quality. To address these issues, this paper proposes an enhanced unsupervised image-to-image translation method based on the Contrastive Unpaired Translation (CUT) model, incorporating Histogram of Oriented Gradients (HOG) features. This novel approach ensures the preservation of the semantic structure of images, even without semantic labels, by minimizing the loss between the HOG features of input and generated images. The method was tested on translating synthetic game environments from GTA5 dataset to realistic urban scenes in cityscapes dataset, demonstrating significant improvements in reducing hallucinations and enhancing image quality.
Authors: Alexander Popov, Alperen Degirmenci, David Wehr, Shashank Hegde, Ryan Oldja, Alexey Kamenev, Bertrand Douillard, David Nist\'er, Urs Muller, Ruchi Bhargava, Stan Birchfield, Nikolai Smolyanskiy
Abstract: We propose the use of latent space generative world models to address the covariate shift problem in autonomous driving. A world model is a neural network capable of predicting an agent's next state given past states and actions. By leveraging a world model during training, the driving policy effectively mitigates covariate shift without requiring an excessive amount of training data. During end-to-end training, our policy learns how to recover from errors by aligning with states observed in human demonstrations, so that at runtime it can recover from perturbations outside the training distribution. Additionally, we introduce a novel transformer-based perception encoder that employs multi-view cross-attention and a learned scene query. We present qualitative and quantitative results, demonstrating significant improvements upon prior state of the art in closed-loop testing in the CARLA simulator, as well as showing the ability to handle perturbations in both CARLA and NVIDIA's DRIVE Sim.