Authors: Maoyu Wang, Yao Lu, Jiaqi Nie, Zeyu Wang, Yun Lin, Qi Xuan, Guan Gui
Abstract: With the rapid development of deep learning, a growing number of pre-trained models have been publicly available. However, deploying these fixed models in real-world IoT applications is challenging because different devices possess heterogeneous computational and memory resources, making it impossible to deploy a single model across all platforms. Although traditional compression methods, such as pruning, quantization, and knowledge distillation, can improve efficiency, they become inflexible once applied and cannot adapt to changing resource constraints. To address these issues, we propose ReStNet, a Reusable and Stitchable Network that dynamically constructs a hybrid network by stitching two pre-trained models together. Implementing ReStNet requires addressing several key challenges, including how to select the optimal stitching points, determine the stitching order of the two pre-trained models, and choose an effective fine-tuning strategy. To systematically address these challenges and adapt to varying resource constraints, ReStNet determines the stitching point by calculating layer-wise similarity via Centered Kernel Alignment (CKA). It then constructs the hybrid model by retaining early layers from a larger-capacity model and appending deeper layers from a smaller one. To facilitate efficient deployment, only the stitching layer is fine-tuned. This design enables rapid adaptation to changing budgets while fully leveraging available resources. Moreover, ReStNet supports both homogeneous (CNN-CNN, Transformer-Transformer) and heterogeneous (CNN-Transformer) stitching, allowing to combine different model families flexibly. Extensive experiments on multiple benchmarks demonstrate that ReStNet achieve flexible accuracy-efficiency trade-offs at runtime while significantly reducing training cost.
Authors: Zhiyu Xue, Reza Abbasi-Asl, Ramtin Pedarsani
Abstract: Generative medical vision-language models~(Med-VLMs) are primarily designed to generate complex textual information~(e.g., diagnostic reports) from multimodal inputs including vision modality~(e.g., medical images) and language modality~(e.g., clinical queries). However, their security vulnerabilities remain underexplored. Med-VLMs should be capable of rejecting harmful queries, such as \textit{Provide detailed instructions for using this CT scan for insurance fraud}. At the same time, addressing security concerns introduces the risk of over-defense, where safety-enhancing mechanisms may degrade general performance, causing Med-VLMs to reject benign clinical queries. In this paper, we propose a novel inference-time defense strategy to mitigate harmful queries, enabling defense against visual and textual jailbreak attacks. Using diverse medical imaging datasets collected from nine modalities, we demonstrate that our defense strategy based on synthetic clinical demonstrations enhances model safety without significantly compromising performance. Additionally, we find that increasing the demonstration budget alleviates the over-defense issue. We then introduce a mixed demonstration strategy as a trade-off solution for balancing security and performance under few-shot demonstration budget constraints.
Authors: Sriram Krishna, Sravan Chittupalli, Sungjae Park
Abstract: In this work, we present BG-HOP, a generative prior that seeks to model bimanual hand-object interactions in 3D. We address the challenge of limited bimanual interaction data by extending existing single-hand generative priors, demonstrating preliminary results in capturing the joint distribution of hands and objects. Our experiments showcase the model's capability to generate bimanual interactions and synthesize grasps for given objects. We make code and models publicly available.
Authors: Peilin Li, Jun Yin, Jing Zhong, Ran Luo, Pengyu Zeng, Miao Zhang
Abstract: In the context of the digital development of architecture, the automatic segmentation of walls and windows is a key step in improving the efficiency of building information models and computer-aided design. This study proposes an automatic segmentation model for building facade walls and windows based on multimodal semantic guidance, called Segment Any Architectural Facades (SAAF). First, SAAF has a multimodal semantic collaborative feature extraction mechanism. By combining natural language processing technology, it can fuse the semantic information in text descriptions with image features, enhancing the semantic understanding of building facade components. Second, we developed an end-to-end training framework that enables the model to autonomously learn the mapping relationship from text descriptions to image segmentation, reducing the influence of manual intervention on the segmentation results and improving the automation and robustness of the model. Finally, we conducted extensive experiments on multiple facade datasets. The segmentation results of SAAF outperformed existing methods in the mIoU metric, indicating that the SAAF model can maintain high-precision segmentation ability when faced with diverse datasets. Our model has made certain progress in improving the accuracy and generalization ability of the wall and window segmentation task. It is expected to provide a reference for the development of architectural computer vision technology and also explore new ideas and technical paths for the application of multimodal learning in the architectural field.
Authors: Xinlong Chen, Yuanxing Zhang, Yushuo Guan, Bohan Zeng, Yang Shi, Sihan Yang, Pengfei Wan, Qiang Liu, Liang Wang, Tieniu Tan
Abstract: Recent advancements in multimodal large language models have successfully extended the Reason-Then-Respond paradigm to image-based reasoning, yet video-based reasoning remains an underdeveloped frontier, primarily due to the scarcity of high-quality reasoning-oriented data and effective training methodologies. To bridge this gap, we introduce DarkEventInfer and MixVidQA, two novel datasets specifically designed to stimulate the model's advanced video understanding and reasoning abilities. DarkEventinfer presents videos with masked event segments, requiring models to infer the obscured content based on contextual video cues. MixVidQA, on the other hand, presents interleaved video sequences composed of two distinct clips, challenging models to isolate and reason about one while disregarding the other. Leveraging these carefully curated training samples together with reinforcement learning guided by diverse reward functions, we develop VersaVid-R1, the first versatile video understanding and reasoning model under the Reason-Then-Respond paradigm capable of handling multiple-choice and open-ended question answering, as well as video captioning tasks. Extensive experiments demonstrate that VersaVid-R1 significantly outperforms existing models across a broad spectrum of benchmarks, covering video general understanding, cognitive reasoning, and captioning tasks.
Authors: Zheqi He, Yesheng Liu, Jing-shu Zheng, Xuejing Li, Richeng Xuan, Jin-Ge Yao, Xi Yang
Abstract: We present FlagEvalMM, an open-source evaluation framework designed to comprehensively assess multimodal models across a diverse range of vision-language understanding and generation tasks, such as visual question answering, text-to-image/video generation, and image-text retrieval. We decouple model inference from evaluation through an independent evaluation service, thus enabling flexible resource allocation and seamless integration of new tasks and models. Moreover, FlagEvalMM utilizes advanced inference acceleration tools (e.g., vLLM, SGLang) and asynchronous data loading to significantly enhance evaluation efficiency. Extensive experiments show that FlagEvalMM offers accurate and efficient insights into model strengths and limitations, making it a valuable tool for advancing multimodal research. The framework is publicly accessible athttps://github.com/flageval-baai/FlagEvalMM.
Authors: Zheda Mai, Arpita Chowdhury, Zihe Wang, Sooyoung Jeon, Lemeng Wang, Jiacheng Hou, Jihyung Kil, Wei-Lun Chao
Abstract: The rise of vision foundation models (VFMs) calls for systematic evaluation. A common approach pairs VFMs with large language models (LLMs) as general-purpose heads, followed by evaluation on broad Visual Question Answering (VQA) benchmarks. However, this protocol has two key blind spots: (i) the instruction tuning data may not align with VQA test distributions, meaning a wrong prediction can stem from such data mismatch rather than a VFM' visual shortcomings; (ii) VQA benchmarks often require multiple visual abilities, making it hard to tell whether errors stem from lacking all required abilities or just a single critical one. To address these gaps, we introduce AVA-Bench, the first benchmark that explicitly disentangles 14 Atomic Visual Abilities (AVAs) -- foundational skills like localization, depth estimation, and spatial understanding that collectively support complex visual reasoning tasks. By decoupling AVAs and matching training and test distributions within each, AVA-Bench pinpoints exactly where a VFM excels or falters. Applying AVA-Bench to leading VFMs thus reveals distinctive "ability fingerprints," turning VFM selection from educated guesswork into principled engineering. Notably, we find that a 0.5B LLM yields similar VFM rankings as a 7B LLM while cutting GPU hours by 8x, enabling more efficient evaluation. By offering a comprehensive and transparent benchmark, we hope AVA-Bench lays the foundation for the next generation of VFMs.
Authors: Jerry Lin, Partick P. W. Chen
Abstract: Accurately labeling (or annotation) data is still a bottleneck in computer vision, especially for large-scale tasks where manual labeling is time-consuming and error-prone. While tools like LabelImg can handle the labeling task, some of them still require annotators to manually label each image. In this paper, we introduce BakuFlow, a streamlining semi-automatic label generation tool. Key features include (1) a live adjustable magnifier for pixel-precise manual corrections, improving user experience; (2) an interactive data augmentation module to diversify training datasets; (3) label propagation for rapidly copying labeled objects between consecutive frames, greatly accelerating annotation of video data; and (4) an automatic labeling module powered by a modified YOLOE framework. Unlike the original YOLOE, our extension supports adding new object classes and any number of visual prompts per class during annotation, enabling flexible and scalable labeling for dynamic, real-world datasets. These innovations make BakuFlow especially effective for object detection and tracking, substantially reducing labeling workload and improving efficiency in practical computer vision and industrial scenarios.
Authors: Xiaofeng Zhang, Michelle Lin, Simon Lacoste-Julien, Aaron Courville, Yash Goyal
Abstract: The widespread adoption of generative AI models has raised growing concerns about representational harm and potential discriminatory outcomes. Yet, despite growing literature on this topic, the mechanisms by which bias emerges - especially in unconditional generation - remain disentangled. We define the bias of an attribute as the difference between the probability of its presence in the observed distribution and its expected proportion in an ideal reference distribution. In our analysis, we train a set of unconditional image generative models and adopt a commonly used bias evaluation framework to study bias shift between training and generated distributions. Our experiments reveal that the detected attribute shifts are small. We find that the attribute shifts are sensitive to the attribute classifier used to label generated images in the evaluation framework, particularly when its decision boundaries fall in high-density regions. Our empirical analysis indicates that this classifier sensitivity is often observed in attributes values that lie on a spectrum, as opposed to exhibiting a binary nature. This highlights the need for more representative labeling practices, understanding the shortcomings through greater scrutiny of evaluation frameworks, and recognizing the socially complex nature of attributes when evaluating bias.
Authors: Arnav Yayavaram, Siddharth Yayavaram, Simran Khanuja, Michael Saxon, Graham Neubig
Abstract: As text-to-image models become increasingly prevalent, ensuring their equitable performance across diverse cultural contexts is critical. Efforts to mitigate cross-cultural biases have been hampered by trade-offs, including a loss in performance, factual inaccuracies, or offensive outputs. Despite widespread recognition of these challenges, an inability to reliably measure these biases has stalled progress. To address this gap, we introduce CAIRe, a novel evaluation metric that assesses the degree of cultural relevance of an image, given a user-defined set of labels. Our framework grounds entities and concepts in the image to a knowledge base and uses factual information to give independent graded judgments for each culture label. On a manually curated dataset of culturally salient but rare items built using language models, CAIRe surpasses all baselines by 28% F1 points. Additionally, we construct two datasets for culturally universal concept, one comprising of T2I-generated outputs and another retrieved from naturally occurring data. CAIRe achieves Pearson's correlations of 0.56 and 0.66 with human ratings on these sets, based on a 5-point Likert scale of cultural relevance. This demonstrates its strong alignment with human judgment across diverse image sources.
Authors: Yu Gao, Haoyuan Guo, Tuyen Hoang, Weilin Huang, Lu Jiang, Fangyuan Kong, Huixia Li, Jiashi Li, Liang Li, Xiaojie Li, Xunsong Li, Yifu Li, Shanchuan Lin, Zhijie Lin, Jiawei Liu, Shu Liu, Xiaonan Nie, Zhiwu Qing, Yuxi Ren, Li Sun, Zhi Tian, Rui Wang, Sen Wang, Guoqiang Wei, Guohong Wu, Jie Wu, Ruiqi Xia, Fei Xiao, Xuefeng Xiao, Jiangqiao Yan, Ceyuan Yang, Jianchao Yang, Runkai Yang, Tao Yang, Yihang Yang, Zilyu Ye, Xuejiao Zeng, Yan Zeng, Heng Zhang, Yang Zhao, Xiaozheng Zheng, Peihao Zhu, Jiaxin Zou, Feilong Zuo
Abstract: Notable breakthroughs in diffusion modeling have propelled rapid improvements in video generation, yet current foundational model still face critical challenges in simultaneously balancing prompt following, motion plausibility, and visual quality. In this report, we introduce Seedance 1.0, a high-performance and inference-efficient video foundation generation model that integrates several core technical improvements: (i) multi-source data curation augmented with precision and meaningful video captioning, enabling comprehensive learning across diverse scenarios; (ii) an efficient architecture design with proposed training paradigm, which allows for natively supporting multi-shot generation and jointly learning of both text-to-video and image-to-video tasks. (iii) carefully-optimized post-training approaches leveraging fine-grained supervised fine-tuning, and video-specific RLHF with multi-dimensional reward mechanisms for comprehensive performance improvements; (iv) excellent model acceleration achieving ~10x inference speedup through multi-stage distillation strategies and system-level optimizations. Seedance 1.0 can generate a 5-second video at 1080p resolution only with 41.4 seconds (NVIDIA-L20). Compared to state-of-the-art video generation models, Seedance 1.0 stands out with high-quality and fast video generation having superior spatiotemporal fluidity with structural stability, precise instruction adherence in complex multi-subject contexts, native multi-shot narrative coherence with consistent subject representation.
Authors: Sungwon Hwang, Hyojin Jang, Kinam Kim, Minho Park, Jaegul choo
Abstract: Fine-tuning Video Diffusion Models (VDMs) at the user level to generate videos that reflect specific attributes of training data presents notable challenges, yet remains underexplored despite its practical importance. Meanwhile, recent work such as Representation Alignment (REPA) has shown promise in improving the convergence and quality of DiT-based image diffusion models by aligning, or assimilating, its internal hidden states with external pretrained visual features, suggesting its potential for VDM fine-tuning. In this work, we first propose a straightforward adaptation of REPA for VDMs and empirically show that, while effective for convergence, it is suboptimal in preserving semantic consistency across frames. To address this limitation, we introduce Cross-frame Representation Alignment (CREPA), a novel regularization technique that aligns hidden states of a frame with external features from neighboring frames. Empirical evaluations on large-scale VDMs, including CogVideoX-5B and Hunyuan Video, demonstrate that CREPA improves both visual fidelity and cross-frame semantic coherence when fine-tuned with parameter-efficient methods such as LoRA. We further validate CREPA across diverse datasets with varying attributes, confirming its broad applicability. Project page: https://crepavideo.github.io
Authors: Mojtaba Nafez, Amirhossein Koochakian, Arad Maleki, Jafar Habibi, Mohammad Hossein Rohban
Abstract: Anomaly Detection (AD) and Anomaly Localization (AL) are crucial in fields that demand high reliability, such as medical imaging and industrial monitoring. However, current AD and AL approaches are often susceptible to adversarial attacks due to limitations in training data, which typically include only normal, unlabeled samples. This study introduces PatchGuard, an adversarially robust AD and AL method that incorporates pseudo anomalies with localization masks within a Vision Transformer (ViT)-based architecture to address these vulnerabilities. We begin by examining the essential properties of pseudo anomalies, and follow it by providing theoretical insights into the attention mechanisms required to enhance the adversarial robustness of AD and AL systems. We then present our approach, which leverages Foreground-Aware Pseudo-Anomalies to overcome the deficiencies of previous anomaly-aware methods. Our method incorporates these crafted pseudo-anomaly samples into a ViT-based framework, with adversarial training guided by a novel loss function designed to improve model robustness, as supported by our theoretical analysis. Experimental results on well-established industrial and medical datasets demonstrate that PatchGuard significantly outperforms previous methods in adversarial settings, achieving performance gains of $53.2\%$ in AD and $68.5\%$ in AL, while also maintaining competitive accuracy in non-adversarial settings. The code repository is available at https://github.com/rohban-lab/PatchGuard .
Authors: Yuchen Zhang, Nikhil Keetha, Chenwei Lyu, Bhuvan Jhamb, Yutian Chen, Yuheng Qiu, Jay Karhade, Shreyas Jha, Yaoyu Hu, Deva Ramanan, Sebastian Scherer, Wenshan Wang
Abstract: Dense image correspondence is central to many applications, such as visual odometry, 3D reconstruction, object association, and re-identification. Historically, dense correspondence has been tackled separately for wide-baseline scenarios and optical flow estimation, despite the common goal of matching content between two images. In this paper, we develop a Unified Flow & Matching model (UFM), which is trained on unified data for pixels that are co-visible in both source and target images. UFM uses a simple, generic transformer architecture that directly regresses the (u,v) flow. It is easier to train and more accurate for large flows compared to the typical coarse-to-fine cost volumes in prior work. UFM is 28% more accurate than state-of-the-art flow methods (Unimatch), while also having 62% less error and 6.7x faster than dense wide-baseline matchers (RoMa). UFM is the first to demonstrate that unified training can outperform specialized approaches across both domains. This result enables fast, general-purpose correspondence and opens new directions for multi-modal, long-range, and real-time correspondence tasks.
Authors: Sindhu Boddu, Arindam Mukherjee
Abstract: This paper presents a lightweight and energy-efficient object detection solution for aerial imagery captured during emergency response situations. We focus on deploying the YOLOv4-Tiny model, a compact convolutional neural network, optimized through post-training quantization to INT8 precision. The model is trained on a custom-curated aerial emergency dataset, consisting of 10,820 annotated images covering critical emergency scenarios. Unlike prior works that rely on publicly available datasets, we created this dataset ourselves due to the lack of publicly available drone-view emergency imagery, making the dataset itself a key contribution of this work. The quantized model is evaluated against YOLOv5-small across multiple metrics, including mean Average Precision (mAP), F1 score, inference time, and model size. Experimental results demonstrate that the quantized YOLOv4-Tiny achieves comparable detection performance while reducing the model size from 22.5 MB to 6.4 MB and improving inference speed by 44\%. With a 71\% reduction in model size and a 44\% increase in inference speed, the quantized YOLOv4-Tiny model proves highly suitable for real-time emergency detection on low-power edge devices.
Authors: Sindhu Boddu, Arindam Mukherjee
Abstract: This paper presents the deployment and performance evaluation of a quantized YOLOv4-Tiny model for real-time object detection in aerial emergency imagery on a resource-constrained edge device the Raspberry Pi 5. The YOLOv4-Tiny model was quantized to INT8 precision using TensorFlow Lite post-training quantization techniques and evaluated for detection speed, power consumption, and thermal feasibility under embedded deployment conditions. The quantized model achieved an inference time of 28.2 ms per image with an average power consumption of 13.85 W, demonstrating a significant reduction in power usage compared to its FP32 counterpart. Detection accuracy remained robust across key emergency classes such as Ambulance, Police, Fire Engine, and Car Crash. These results highlight the potential of low-power embedded AI systems for real-time deployment in safety-critical emergency response applications.
Authors: Tong Wang, Guanzhou Chen, Xiaodong Zhang, Chenxi Liu, Jiaqi Wang, Xiaoliang Tan, Wenchao Guo, Qingyuan Yang, Kaiqi Zhang
Abstract: Remote sensing image interpretation plays a critical role in environmental monitoring, urban planning, and disaster assessment. However, acquiring high-quality labeled data is often costly and time-consuming. To address this challenge, we proposes a multi-modal self-supervised learning framework that leverages high-resolution RGB images, multi-spectral data, and digital surface models (DSM) for pre-training. By designing an information-aware adaptive masking strategy, cross-modal masking mechanism, and multi-task self-supervised objectives, the framework effectively captures both the correlations across different modalities and the unique feature structures within each modality. We evaluated the proposed method on multiple downstream tasks, covering typical remote sensing applications such as scene classification, semantic segmentation, change detection, object detection, and depth estimation. Experiments are conducted on 15 remote sensing datasets, encompassing 26 tasks. The results demonstrate that the proposed method outperforms existing pretraining approaches in most tasks. Specifically, on the Potsdam and Vaihingen semantic segmentation tasks, our method achieved mIoU scores of 78.30\% and 76.50\%, with only 50\% train-set. For the US3D depth estimation task, the RMSE error is reduced to 0.182, and for the binary change detection task in SECOND dataset, our method achieved mIoU scores of 47.51\%, surpassing the second CS-MAE by 3 percentage points. Our pretrain code, checkpoints, and HR-Pairs dataset can be found in https://github.com/CVEO/MSSDF.
Authors: Yuxing Long, Jiyao Zhang, Mingjie Pan, Tianshu Wu, Taewhan Kim, Hao Dong
Abstract: Correct use of electrical appliances has significantly improved human life quality. Unlike simple tools that can be manipulated with common sense, different parts of electrical appliances have specific functions defined by manufacturers. If we want the robot to heat bread by microwave, we should enable them to review the microwave manual first. From the manual, it can learn about component functions, interaction methods, and representative task steps about appliances. However, previous manual-related works remain limited to question-answering tasks while existing manipulation researchers ignore the manual's important role and fail to comprehend multi-page manuals. In this paper, we propose the first manual-based appliance manipulation benchmark CheckManual. Specifically, we design a large model-assisted human-revised data generation pipeline to create manuals based on CAD appliance models. With these manuals, we establish novel manual-based manipulation challenges, metrics, and simulator environments for model performance evaluation. Furthermore, we propose the first manual-based manipulation planning model ManualPlan to set up a group of baselines for the CheckManual benchmark.
Authors: Songping Wang, Xiantao Hu, Yueming Lyu, Caifeng Shan
Abstract: Recently, multimodal tasks have strongly advanced the field of action recognition with their rich multimodal information. However, due to the scarcity of tri-modal data, research on tri-modal action recognition tasks faces many challenges. To this end, we have proposed a comprehensive multimodal action recognition solution that effectively utilizes multimodal information. First, the existing data are transformed and expanded by optimizing data enhancement techniques to enlarge the training scale. At the same time, more RGB datasets are used to pre-train the backbone network, which is better adapted to the new task by means of transfer learning. Secondly, multimodal spatial features are extracted with the help of 2D CNNs and combined with the Temporal Shift Module (TSM) to achieve multimodal spatial-temporal feature extraction comparable to 3D CNNs and improve the computational efficiency. In addition, common prediction enhancement methods, such as Stochastic Weight Averaging (SWA), Ensemble and Test-Time augmentation (TTA), are used to integrate the knowledge of models from different training periods of the same architecture and different architectures, so as to predict the actions from different perspectives and fully exploit the target information. Ultimately, we achieved the Top-1 accuracy of 99% and the Top-5 accuracy of 100% on the competition leaderboard, demonstrating the superiority of our solution.
Authors: Shanchuan Lin, Ceyuan Yang, Hao He, Jianwen Jiang, Yuxi Ren, Xin Xia, Yang Zhao, Xuefeng Xiao, Lu Jiang
Abstract: Existing large-scale video generation models are computationally intensive, preventing adoption in real-time and interactive applications. In this work, we propose autoregressive adversarial post-training (AAPT) to transform a pre-trained latent video diffusion model into a real-time, interactive video generator. Our model autoregressively generates a latent frame at a time using a single neural function evaluation (1NFE). The model can stream the result to the user in real time and receive interactive responses as controls to generate the next latent frame. Unlike existing approaches, our method explores adversarial training as an effective paradigm for autoregressive generation. This not only allows us to design an architecture that is more efficient for one-step generation while fully utilizing the KV cache, but also enables training the model in a student-forcing manner that proves to be effective in reducing error accumulation during long video generation. Our experiments demonstrate that our 8B model achieves real-time, 24fps, streaming video generation at 736x416 resolution on a single H100, or 1280x720 on 8xH100 up to a minute long (1440 frames). Visit our research website at https://seaweed-apt.com/2
Authors: Junchao Zhou
Abstract: Image segmentation is a fundamental task in computer vision aimed at delineating object boundaries within images. Traditional approaches, such as edge detection and variational methods, have been widely explored, while recent advances in deep learning have shown promising results but often require extensive training data. In this work, we propose a novel variational framework for 2D image segmentation that integrates concepts from shape analysis and diffeomorphic transformations. Our method models segmentation as the deformation of a template curve via a diffeomorphic transformation of the image domain, using the Large Deformation Diffeomorphic Metric Mapping (LDDMM) framework. The curve evolution is guided by a loss function that compares the deformed curve to the image gradient field, formulated through the varifold representation of geometric shapes. The approach is implemented in Python with GPU acceleration using the PyKeops library. This framework allows for accurate segmentation with a flexible and theoretically grounded methodology that does not rely on large datasets.
Authors: Hongguang Zhu, Yunchao Wei, Mengyu Wang, Siyu Jiao, Yan Fang, Jiannan Huang, Yao Zhao
Abstract: Diffusion models (DMs) have achieved significant progress in text-to-image generation. However, the inevitable inclusion of sensitive information during pre-training poses safety risks, such as unsafe content generation and copyright infringement. Concept erasing finetunes weights to unlearn undesirable concepts, and has emerged as a promising solution. However, existing methods treat unsafe concept as a fixed word and repeatedly erase it, trapping DMs in ``word concept abyss'', which prevents generalized concept-related erasing. To escape this abyss, we introduce semantic-augment erasing which transforms concept word erasure into concept domain erasure by the cyclic self-check and self-erasure. It efficiently explores and unlearns the boundary representation of concept domain through semantic spatial relationships between original and training DMs, without requiring additional preprocessed data. Meanwhile, to mitigate the retention degradation of irrelevant concepts while erasing unsafe concepts, we further propose the global-local collaborative retention mechanism that combines global semantic relationship alignment with local predicted noise preservation, effectively expanding the retentive receptive field for irrelevant concepts. We name our method SAGE, and extensive experiments demonstrate the comprehensive superiority of SAGE compared with other methods in the safe generation of DMs. The code and weights will be open-sourced at https://github.com/KevinLight831/SAGE.
Authors: Zeran Ke, Bin Tan, Xianwei Zheng, Yujun Shen, Tianfu Wu, Nan Xue
Abstract: This paper studies the problem of Line Segment Detection (LSD) for the characterization of line geometry in images, with the aim of learning a domain-agnostic robust LSD model that works well for any natural images. With the focus of scalable self-supervised learning of LSD, we revisit and streamline the fundamental designs of (deep and non-deep) LSD approaches to have a high-performing and efficient LSD learner, dubbed as ScaleLSD, for the curation of line geometry at scale from over 10M unlabeled real-world images. Our ScaleLSD works very well to detect much more number of line segments from any natural images even than the pioneered non-deep LSD approach, having a more complete and accurate geometric characterization of images using line segments. Experimentally, our proposed ScaleLSD is comprehensively testified under zero-shot protocols in detection performance, single-view 3D geometry estimation, two-view line segment matching, and multiview 3D line mapping, all with excellent performance obtained. Based on the thorough evaluation, our ScaleLSD is observed to be the first deep approach that outperforms the pioneered non-deep LSD in all aspects we have tested, significantly expanding and reinforcing the versatility of the line geometry of images. Code and Models are available at https://github.com/ant-research/scalelsd
Authors: Qijian Tian, Xin Tan, Jingyu Gong, Yuan Xie, Lizhuang Ma
Abstract: We propose a feed-forward Gaussian Splatting model that unifies 3D scene and semantic field reconstruction. Combining 3D scenes with semantic fields facilitates the perception and understanding of the surrounding environment. However, key challenges include embedding semantics into 3D representations, achieving generalizable real-time reconstruction, and ensuring practical applicability by using only images as input without camera parameters or ground truth depth. To this end, we propose UniForward, a feed-forward model to predict 3D Gaussians with anisotropic semantic features from only uncalibrated and unposed sparse-view images. To enable the unified representation of the 3D scene and semantic field, we embed semantic features into 3D Gaussians and predict them through a dual-branch decoupled decoder. During training, we propose a loss-guided view sampler to sample views from easy to hard, eliminating the need for ground truth depth or masks required by previous methods and stabilizing the training process. The whole model can be trained end-to-end using a photometric loss and a distillation loss that leverages semantic features from a pre-trained 2D semantic model. At the inference stage, our UniForward can reconstruct 3D scenes and the corresponding semantic fields in real time from only sparse-view images. The reconstructed 3D scenes achieve high-quality rendering, and the reconstructed 3D semantic field enables the rendering of view-consistent semantic features from arbitrary views, which can be further decoded into dense segmentation masks in an open-vocabulary manner. Experiments on novel view synthesis and novel view segmentation demonstrate that our method achieves state-of-the-art performances for unifying 3D scene and semantic field reconstruction.
Authors: Jialong Zuo, Yongtai Deng, Mengdan Tan, Rui Jin, Dongyue Wu, Nong Sang, Liang Pan, Changxin Gao
Abstract: In real-word scenarios, person re-identification (ReID) expects to identify a person-of-interest via the descriptive query, regardless of whether the query is a single modality or a combination of multiple modalities. However, existing methods and datasets remain constrained to limited modalities, failing to meet this requirement. Therefore, we investigate a new challenging problem called Omni Multi-modal Person Re-identification (OM-ReID), which aims to achieve effective retrieval with varying multi-modal queries. To address dataset scarcity, we construct ORBench, the first high-quality multi-modal dataset comprising 1,000 unique identities across five modalities: RGB, infrared, color pencil, sketch, and textual description. This dataset also has significant superiority in terms of diversity, such as the painting perspectives and textual information. It could serve as an ideal platform for follow-up investigations in OM-ReID. Moreover, we propose ReID5o, a novel multi-modal learning framework for person ReID. It enables synergistic fusion and cross-modal alignment of arbitrary modality combinations in a single model, with a unified encoding and multi-expert routing mechanism proposed. Extensive experiments verify the advancement and practicality of our ORBench. A wide range of possible models have been evaluated and compared on it, and our proposed ReID5o model gives the best performance. The dataset and code will be made publicly available at https://github.com/Zplusdragon/ReID5o_ORBench.
Authors: Kaiyu Guo, Zijian Wang, Brian C. Lovell, Mahsa Baktashmotlagh
Abstract: Out-of-Distribution (OOD) detection is essential for the trustworthiness of AI systems. Methods using prior information (i.e., subspace-based methods) have shown effective performance by extracting information geometry to detect OOD data with a more appropriate distance metric. However, these methods fail to address the geometry distorted by ill-distributed samples, due to the limitation of statically extracting information geometry from the training distribution. In this paper, we argue that the influence of ill-distributed samples can be corrected by dynamically adjusting the prior geometry in response to new data. Based on this insight, we propose a novel approach that dynamically updates the prior covariance matrix using real-time input features, refining its information. Specifically, we reduce the covariance along the direction of real-time input features and constrain adjustments to the residual space, thus preserving essential data characteristics and avoiding effects on unintended directions in the principal space. We evaluate our method on two pre-trained models for the CIFAR dataset and five pre-trained models for ImageNet-1k, including the self-supervised DINO model. Extensive experiments demonstrate that our approach significantly enhances OOD detection across various models. The code is released at https://github.com/workerbcd/ooddcc.
Authors: Xinya Liu, Jianghao Wu, Tao Lu, Shaoting Zhang, Guotai Wang
Abstract: Domain Adaptation (DA) is crucial for robust deployment of medical image segmentation models when applied to new clinical centers with significant domain shifts. Source-Free Domain Adaptation (SFDA) is appealing as it can deal with privacy concerns and access constraints on source-domain data during adaptation to target-domain data. However, SFDA faces challenges such as insufficient supervision in the target domain with unlabeled images. In this work, we propose a Segment Anything Model (SAM)-guided Reliable Pseudo-Labels method for SFDA (SRPL-SFDA) with three key components: 1) Test-Time Tri-branch Intensity Enhancement (T3IE) that not only improves quality of raw pseudo-labels in the target domain, but also leads to SAM-compatible inputs with three channels to better leverage SAM's zero-shot inference ability for refining the pseudo-labels; 2) A reliable pseudo-label selection module that rejects low-quality pseudo-labels based on Consistency of Multiple SAM Outputs (CMSO) under input perturbations with T3IE; and 3) A reliability-aware training procedure in the unlabeled target domain where reliable pseudo-labels are used for supervision and unreliable parts are regularized by entropy minimization. Experiments conducted on two multi-domain medical image segmentation datasets for fetal brain and the prostate respectively demonstrate that: 1) SRPL-SFDA effectively enhances pseudo-label quality in the unlabeled target domain, and improves SFDA performance by leveraging the reliability-aware training; 2) SRPL-SFDA outperformed state-of-the-art SFDA methods, and its performance is close to that of supervised training in the target domain. The code of this work is available online: https://github.com/HiLab-git/SRPL-SFDA.
Authors: Vaclav Knapp, Matyas Bohacek
Abstract: In video understanding tasks, particularly those involving human motion, synthetic data generation often suffers from uncanny features, diminishing its effectiveness for training. Tasks such as sign language translation, gesture recognition, and human motion understanding in autonomous driving have thus been unable to exploit the full potential of synthetic data. This paper proposes a method for generating synthetic human action video data using pose transfer (specifically, controllable 3D Gaussian avatar models). We evaluate this method on the Toyota Smarthome and NTU RGB+D datasets and show that it improves performance in action recognition tasks. Moreover, we demonstrate that the method can effectively scale few-shot datasets, making up for groups underrepresented in the real training data and adding diverse backgrounds. We open-source the method along with RANDOM People, a dataset with videos and avatars of novel human identities for pose transfer crowd-sourced from the internet.
Authors: Xinyu Peng, Ziyang Zheng, Yaoming Wang, Han Li, Nuowen Kan, Wenrui Dai, Chenglin Li, Junni Zou, Hongkai Xiong
Abstract: We propose Noise Conditional Variational Score Distillation (NCVSD), a novel method for distilling pretrained diffusion models into generative denoisers. We achieve this by revealing that the unconditional score function implicitly characterizes the score function of denoising posterior distributions. By integrating this insight into the Variational Score Distillation (VSD) framework, we enable scalable learning of generative denoisers capable of approximating samples from the denoising posterior distribution across a wide range of noise levels. The proposed generative denoisers exhibit desirable properties that allow fast generation while preserve the benefit of iterative refinement: (1) fast one-step generation through sampling from pure Gaussian noise at high noise levels; (2) improved sample quality by scaling the test-time compute with multi-step sampling; and (3) zero-shot probabilistic inference for flexible and controllable sampling. We evaluate NCVSD through extensive experiments, including class-conditional image generation and inverse problem solving. By scaling the test-time compute, our method outperforms teacher diffusion models and is on par with consistency models of larger sizes. Additionally, with significantly fewer NFEs than diffusion-based methods, we achieve record-breaking LPIPS on inverse problems.
Authors: Yunxiao Shi, Yinhao Zhu, Shizhong Han, Jisoo Jeong, Amin Ansari, Hong Cai, Fatih Porikli
Abstract: Occupancy prediction infers fine-grained 3D geometry and semantics from camera images of the surrounding environment, making it a critical perception task for autonomous driving. Existing methods either adopt dense grids as scene representation, which is difficult to scale to high resolution, or learn the entire scene using a single set of sparse queries, which is insufficient to handle the various object characteristics. In this paper, we present ODG, a hierarchical dual sparse Gaussian representation to effectively capture complex scene dynamics. Building upon the observation that driving scenes can be universally decomposed into static and dynamic counterparts, we define dual Gaussian queries to better model the diverse scene objects. We utilize a hierarchical Gaussian transformer to predict the occupied voxel centers and semantic classes along with the Gaussian parameters. Leveraging the real-time rendering capability of 3D Gaussian Splatting, we also impose rendering supervision with available depth and semantic map annotations injecting pixel-level alignment to boost occupancy learning. Extensive experiments on the Occ3D-nuScenes and Occ3D-Waymo benchmarks demonstrate our proposed method sets new state-of-the-art results while maintaining low inference cost.
Authors: Yukang Feng, Jianwen Sun, Chuanhao Li, Zizhen Li, Jiaxin Ai, Fanrui Zhang, Yifan Chang, Sizhuo Zhou, Shenglin Zhang, Yu Dai, Kaipeng Zhang
Abstract: Recent advancements in Large Multimodal Models (LMMs) have significantly improved multimodal understanding and generation. However, these models still struggle to generate tightly interleaved image-text outputs, primarily due to the limited scale, quality and instructional richness of current training datasets. To address this, we introduce InterSyn, a large-scale multimodal dataset constructed using our Self-Evaluation with Iterative Refinement (SEIR) method. InterSyn features multi-turn, instruction-driven dialogues with tightly interleaved imagetext responses, providing rich object diversity and rigorous automated quality refinement, making it well-suited for training next-generation instruction-following LMMs. Furthermore, to address the lack of reliable evaluation tools capable of assessing interleaved multimodal outputs, we introduce SynJudge, an automatic evaluation model designed to quantitatively assess multimodal outputs along four dimensions: text content, image content, image quality, and image-text synergy. Experimental studies show that the SEIR method leads to substantially higher dataset quality compared to an otherwise identical process without refinement. Moreover, LMMs trained on InterSyn achieve uniform performance gains across all evaluation metrics, confirming InterSyn's utility for advancing multimodal systems.
Authors: Swadhin Das, Divyansh Mundra, Priyanshu Dayal, Raksha Sharma
Abstract: Transformer-based models have achieved strong performance in remote sensing image captioning by capturing long-range dependencies and contextual information. However, their practical deployment is hindered by high computational costs, especially in multi-modal frameworks that employ separate transformer-based encoders and decoders. In addition, existing remote sensing image captioning models primarily focus on high-level semantic extraction while often overlooking fine-grained structural features such as edges, contours, and object boundaries. To address these challenges, a lightweight transformer architecture is proposed by reducing the dimensionality of the encoder layers and employing a distilled version of GPT-2 as the decoder. A knowledge distillation strategy is used to transfer knowledge from a more complex teacher model to improve the performance of the lightweight network. Furthermore, an edge-aware enhancement strategy is incorporated to enhance image representation and object boundary understanding, enabling the model to capture fine-grained spatial details in remote sensing images. Experimental results demonstrate that the proposed approach significantly improves caption quality compared to state-of-the-art methods.
Authors: Ayush Gupta, Anirban Roy, Rama Chellappa, Nathaniel D. Bastian, Alvaro Velasquez, Susmit Jha
Abstract: We address the problem of video question answering (video QA) with temporal grounding in a weakly supervised setup, without any temporal annotations. Given a video and a question, we generate an open-ended answer grounded with the start and end time. For this task, we propose TOGA: a vision-language model for Temporally Grounded Open-Ended Video QA with Weak Supervision. We instruct-tune TOGA to jointly generate the answer and the temporal grounding. We operate in a weakly supervised setup where the temporal grounding annotations are not available. We generate pseudo labels for temporal grounding and ensure the validity of these labels by imposing a consistency constraint between the question of a grounding response and the response generated by a question referring to the same temporal segment. We notice that jointly generating the answers with the grounding improves performance on question answering as well as grounding. We evaluate TOGA on grounded QA and open-ended QA tasks. For grounded QA, we consider the NExT-GQA benchmark which is designed to evaluate weakly supervised grounded question answering. For open-ended QA, we consider the MSVD-QA and ActivityNet-QA benchmarks. We achieve state-of-the-art performance for both tasks on these benchmarks.
Authors: Yuhe Ding, Jian Liang, Bo Jiang, Zi Wang, Aihua Zheng, Bin Luo
Abstract: CLIP-based domain generalization aims to improve model generalization to unseen domains by leveraging the powerful zero-shot classification capabilities of CLIP and multiple source datasets. Existing methods typically train a single model across multiple source domains to capture domain-shared information. However, this paradigm inherently suffers from two types of conflicts: 1) sample conflicts, arising from noisy samples and extreme domain shifts among sources; and 2) optimization conflicts, stemming from competition and trade-offs during multi-source training. Both hinder the generalization and lead to suboptimal solutions. Recent studies have shown that model merging can effectively mitigate the competition of multi-objective optimization and improve generalization performance. Inspired by these findings, we propose Harmonizing and Merging (HAM), a novel source model merging framework for CLIP-based domain generalization. During the training process of the source models, HAM enriches the source samples without conflicting samples, and harmonizes the update directions of all models. Then, a redundancy-aware historical model merging method is introduced to effectively integrate knowledge across all source models. HAM comprehensively consolidates source domain information while enabling mutual enhancement among source models, ultimately yielding a final model with optimal generalization capabilities. Extensive experiments on five widely used benchmark datasets demonstrate the effectiveness of our approach, achieving state-of-the-art performance.
Authors: Amirreza Khoshbakht, Erchan Aptoula
Abstract: Open-set domain generalization(OSDG) for hyperspectral image classification presents significant challenges due to the presence of unknown classes in target domains and the need for models to generalize across multiple unseen domains without target-specific adaptation. Existing domain adaptation methods assume access to target domain data during training and fail to address the fundamental issue of domain shift when unknown classes are present, leading to negative transfer and reduced classification performance. To address these limitations, we propose a novel open-set domain generalization framework that combines four key components: Spectrum-Invariant Frequency Disentanglement (SIFD) for domain-agnostic feature extraction, Dual-Channel Residual Network (DCRN) for robust spectral-spatial feature learning, Evidential Deep Learning (EDL) for uncertainty quantification, and Spectral-Spatial Uncertainty Disentanglement (SSUD) for reliable open-set classification. The SIFD module extracts domain-invariant spectral features in the frequency domain through attention-weighted frequency analysis and domain-agnostic regularization, while DCRN captures complementary spectral and spatial information via parallel pathways with adaptive fusion. EDL provides principled uncertainty estimation using Dirichlet distributions, enabling the SSUD module to make reliable open-set decisions through uncertainty-aware pathway weighting and adaptive rejection thresholding. Experimental results on three cross-scene hyperspectral classification tasks show that our approach achieves performance comparable to state-of-the-art domain adaptation methods while requiring no access to the target domain during training. The implementation will be made available at https://github.com/amir-khb/SSUDOSDG upon acceptance.
Authors: Maria Damanaki, Nikos Piperigkos, Alexandros Gkillas, Aris S. Lalos
Abstract: Multi-Object Tracking (MOT) plays a crucial role in autonomous driving systems, as it lays the foundations for advanced perception and precise path planning modules. Nonetheless, single agent based MOT lacks in sensing surroundings due to occlusions, sensors failures, etc. Hence, the integration of multiagent information is essential for comprehensive understanding of the environment. This paper proposes a novel Cooperative MOT framework for tracking objects in 3D LiDAR scene by formulating and solving a graph topology-aware optimization problem so as to fuse information coming from multiple vehicles. By exploiting a fully connected graph topology defined by the detected bounding boxes, we employ the Graph Laplacian processing optimization technique to smooth the position error of bounding boxes and effectively combine them. In that manner, we reveal and leverage inherent coherences of diverse multi-agent detections, and associate the refined bounding boxes to tracked objects at two stages, optimizing localization and tracking accuracies. An extensive evaluation study has been conducted, using the real-world V2V4Real dataset, where the proposed method significantly outperforms the baseline frameworks, including the state-of-the-art deep-learning DMSTrack and V2V4Real, in various testing sequences.
Authors: Cheng Chen, Yunpeng Zhai, Yifan Zhao, Jinyang Gao, Bolin Ding, Jia Li
Abstract: In-context learning (ICL), a predominant trend in instruction learning, aims at enhancing the performance of large language models by providing clear task guidance and examples, improving their capability in task understanding and execution. This paper investigates ICL on Large Vision-Language Models (LVLMs) and explores the policies of multi-modal demonstration selection. Existing research efforts in ICL face significant challenges: First, they rely on pre-defined demonstrations or heuristic selecting strategies based on human intuition, which are usually inadequate for covering diverse task requirements, leading to sub-optimal solutions; Second, individually selecting each demonstration fails in modeling the interactions between them, resulting in information redundancy. Unlike these prevailing efforts, we propose a new exploration-exploitation reinforcement learning framework, which explores policies to fuse multi-modal information and adaptively select adequate demonstrations as an integrated whole. The framework allows LVLMs to optimize themselves by continually refining their demonstrations through self-exploration, enabling the ability to autonomously identify and generate the most effective selection policies for in-context learning. Experimental results verify the superior performance of our approach on four Visual Question-Answering (VQA) datasets, demonstrating its effectiveness in enhancing the generalization capability of few-shot LVLMs.
Authors: Tianxiang Hao, Lixian Zhang, Yingjia Zhang, Mengxuan Chen, Jinxiao Zhang, Haohuan Fu
Abstract: Historical satellite imagery, such as mid-20$^{th}$ century Keyhole data, offers rare insights into understanding early urban development and long-term transformation. However, severe quality degradation (e.g., distortion, misalignment, and spectral scarcity) and annotation absence have long hindered semantic segmentation on such historical RS imagery. To bridge this gap and enhance understanding of urban development, we introduce $\textbf{Urban1960SatBench}$, an annotated segmentation dataset based on historical satellite imagery with the earliest observation time among all existing segmentation datasets, along with a benchmark framework for unsupervised segmentation tasks, $\textbf{Urban1960SatUSM}$. First, $\textbf{Urban1960SatBench}$ serves as a novel, expertly annotated semantic segmentation dataset built on mid-20$^{th}$ century Keyhole imagery, covering 1,240 km$^2$ and key urban classes (buildings, roads, farmland, water). As the earliest segmentation dataset of its kind, it provides a pioneering benchmark for historical urban understanding. Second, $\textbf{Urban1960SatUSM}$(Unsupervised Segmentation Model) is a novel unsupervised semantic segmentation framework for historical RS imagery. It employs a confidence-aware alignment mechanism and focal-confidence loss based on a self-supervised learning architecture, which generates robust pseudo-labels and adaptively prioritizes prediction difficulty and label reliability to improve unsupervised segmentation on noisy historical data without manual supervision. Experiments show Urban1960SatUSM significantly outperforms existing unsupervised segmentation methods on Urban1960SatSeg for segmenting historical urban scenes, promising in paving the way for quantitative studies of long-term urban change using modern computer vision. Our benchmark and supplementary material are available at https://github.com/Tianxiang-Hao/Urban1960SatSeg.
Authors: Zetian Song, Jiaye Fu, Jiaqi Zhang, Xiaohan Lu, Chuanmin Jia, Siwei Ma, Wen Gao
Abstract: The recent development of feedforward 3D Gaussian Splatting (3DGS) presents a new paradigm to reconstruct 3D scenes. Using neural networks trained on large-scale multi-view datasets, it can directly infer 3DGS representations from sparse input views. Although the feedforward approach achieves high reconstruction speed, it still suffers from the substantial storage cost of 3D Gaussians. Existing 3DGS compression methods relying on scene-wise optimization are not applicable due to architectural incompatibilities. To overcome this limitation, we propose TinySplat, a complete feedforward approach for generating compact 3D scene representations. Built upon standard feedforward 3DGS methods, TinySplat integrates a training-free compression framework that systematically eliminates key sources of redundancy. Specifically, we introduce View-Projection Transformation (VPT) to reduce geometric redundancy by projecting geometric parameters into a more compact space. We further present Visibility-Aware Basis Reduction (VABR), which mitigates perceptual redundancy by aligning feature energy along dominant viewing directions via basis transformation. Lastly, spatial redundancy is addressed through an off-the-shelf video codec. Comprehensive experimental results on multiple benchmark datasets demonstrate that TinySplat achieves over 100x compression for 3D Gaussian data generated by feedforward methods. Compared to the state-of-the-art compression approach, we achieve comparable quality with only 6% of the storage size. Meanwhile, our compression framework requires only 25% of the encoding time and 1% of the decoding time.
Authors: Dingcheng Zhen, Qian Qiao, Tan Yu, Kangxi Wu, Ziwei Zhang, Siyuan Liu, Shunshun Yin, Ming Tao
Abstract: We introduce TransDiff, the first image generation model that marries Autoregressive (AR) Transformer with diffusion models. In this joint modeling framework, TransDiff encodes labels and images into high-level semantic features and employs a diffusion model to estimate the distribution of image samples. On the ImageNet 256x256 benchmark, TransDiff significantly outperforms other image generation models based on standalone AR Transformer or diffusion models. Specifically, TransDiff achieves a Fr\'echet Inception Distance (FID) of 1.61 and an Inception Score (IS) of 293.4, and further provides x2 faster inference latency compared to state-of-the-art methods based on AR Transformer and x112 faster inference compared to diffusion-only models. Furthermore, building on the TransDiff model, we introduce a novel image generation paradigm called Multi-Reference Autoregression (MRAR), which performs autoregressive generation by predicting the next image. MRAR enables the model to reference multiple previously generated images, thereby facilitating the learning of more diverse representations and improving the quality of generated images in subsequent iterations. By applying MRAR, the performance of TransDiff is improved, with the FID reduced from 1.61 to 1.42. We expect TransDiff to open up a new frontier in the field of image generation.
Authors: Changhao Peng, Yuqi Ye, Wei Gao
Abstract: Gaussian and Laplacian entropy models are proved effective in learned point cloud attribute compression, as they assist in arithmetic coding of latents. However, we demonstrate through experiments that there is still unutilized information in entropy parameters estimated by neural networks in current methods, which can be used for more accurate probability estimation. Thus we introduce generalized Gaussian entropy model, which controls the tail shape through shape parameter to more accurately estimate the probability of latents. Meanwhile, to the best of our knowledge, existing methods use fixed likelihood intervals for each integer during arithmetic coding, which limits model performance. We propose Mean Error Discriminator (MED) to determine whether the entropy parameter estimation is accurate and then dynamically adjust likelihood intervals. Experiments show that our method significantly improves rate-distortion (RD) performance on three VAE-based models for point cloud attribute compression, and our method can be applied to other compression tasks, such as image and video compression.
Authors: Jianing Chen, Zehao Li, Yujun Cai, Hao Jiang, Chengxuan Qian, Juyuan Kang, Shuqin Gao, Honglong Zhao, Tianlu Mao, Yucheng Zhang
Abstract: Reconstructing dynamic 3D scenes from monocular videos remains a fundamental challenge in 3D vision. While 3D Gaussian Splatting (3DGS) achieves real-time rendering in static settings, extending it to dynamic scenes is challenging due to the difficulty of learning structured and temporally consistent motion representations. This challenge often manifests as three limitations in existing methods: redundant Gaussian updates, insufficient motion supervision, and weak modeling of complex non-rigid deformations. These issues collectively hinder coherent and efficient dynamic reconstruction. To address these limitations, we propose HAIF-GS, a unified framework that enables structured and consistent dynamic modeling through sparse anchor-driven deformation. It first identifies motion-relevant regions via an Anchor Filter to suppresses redundant updates in static areas. A self-supervised Induced Flow-Guided Deformation module induces anchor motion using multi-frame feature aggregation, eliminating the need for explicit flow labels. To further handle fine-grained deformations, a Hierarchical Anchor Propagation mechanism increases anchor resolution based on motion complexity and propagates multi-level transformations. Extensive experiments on synthetic and real-world benchmarks validate that HAIF-GS significantly outperforms prior dynamic 3DGS methods in rendering quality, temporal coherence, and reconstruction efficiency.
Authors: Beomsik Cho, Jaehyung Kim
Abstract: Large Vision-Language Models (LVLMs) have demonstrated remarkable performance across various multimodal tasks by integrating visual perception with language understanding. However, conventional decoding strategies of LVLMs often fail to successfully utilize visual information, leading to visually ungrounded responses. While various approaches have been proposed to address this limitation, they typically require additional training, multi-step inference procedures, or external model dependencies. This paper introduces ReVisiT, a simple yet effective decoding method that references vision tokens to guide the text generation process in LVLMs. Our approach leverages the semantic information embedded within vision tokens by projecting them into the text token distribution space, and dynamically selecting the most relevant vision token at each decoding step through constrained divergence minimization. This selected vision token is then used to refine the output distribution to better incorporate visual semantics. Experiments on three LVLM hallucination benchmarks with two recent LVLMs demonstrate that ReVisiT consistently enhances visual grounding with minimal computational overhead. Moreover, our method achieves competitive or superior results relative to state-of-the-art baselines while reducing computational costs for up to $2\times$.
Authors: Tao Wang, Mengyu Li, Geduo Zeng, Cheng Meng, Qiong Zhang
Abstract: 3D Gaussian Splatting (3DGS) has emerged as a powerful technique for radiance field rendering, but it typically requires millions of redundant Gaussian primitives, overwhelming memory and rendering budgets. Existing compaction approaches address this by pruning Gaussians based on heuristic importance scores, without global fidelity guarantee. To bridge this gap, we propose a novel optimal transport perspective that casts 3DGS compaction as global Gaussian mixture reduction. Specifically, we first minimize the composite transport divergence over a KD-tree partition to produce a compact geometric representation, and then decouple appearance from geometry by fine-tuning color and opacity attributes with far fewer Gaussian primitives. Experiments on benchmark datasets show that our method (i) yields negligible loss in rendering quality (PSNR, SSIM, LPIPS) compared to vanilla 3DGS with only 10% Gaussians; and (ii) consistently outperforms state-of-the-art 3DGS compaction techniques. Notably, our method is applicable to any stage of vanilla or accelerated 3DGS pipelines, providing an efficient and agnostic pathway to lightweight neural rendering.
Authors: Wenjun Ji, Yuxiang Fu, Luyang Ying, Deng-Ping Fan, Yuyi Wang, Ming-Ming Cheng, Ivor Tsang, Qing Guo
Abstract: Cutting-edge works have demonstrated that text-to-image (T2I) diffusion models can generate adversarial patches that mislead state-of-the-art object detectors in the physical world, revealing detectors' vulnerabilities and risks. However, these methods neglect the T2I patches' attack effectiveness when observed from different views in the physical world (i.e., angle robustness of the T2I adversarial patches). In this paper, we study the angle robustness of T2I adversarial patches comprehensively, revealing their angle-robust issues, demonstrating that texts affect the angle robustness of generated patches significantly, and task-specific linguistic instructions fail to enhance the angle robustness. Motivated by the studies, we introduce Angle-Robust Concept Learning (AngleRoCL), a simple and flexible approach that learns a generalizable concept (i.e., text embeddings in implementation) representing the capability of generating angle-robust patches. The learned concept can be incorporated into textual prompts and guides T2I models to generate patches with their attack effectiveness inherently resistant to viewpoint variations. Through extensive simulation and physical-world experiments on five SOTA detectors across multiple views, we demonstrate that AngleRoCL significantly enhances the angle robustness of T2I adversarial patches compared to baseline methods. Our patches maintain high attack success rates even under challenging viewing conditions, with over 50% average relative improvement in attack effectiveness across multiple angles. This research advances the understanding of physically angle-robust patches and provides insights into the relationship between textual concepts and physical properties in T2I-generated contents.
Authors: Yi Zhang, Yi Wang, Yawen Cui, Lap-Pui Chau
Abstract: This paper proposes 3DGeoDet, a novel geometry-aware 3D object detection approach that effectively handles single- and multi-view RGB images in indoor and outdoor environments, showcasing its general-purpose applicability. The key challenge for image-based 3D object detection tasks is the lack of 3D geometric cues, which leads to ambiguity in establishing correspondences between images and 3D representations. To tackle this problem, 3DGeoDet generates efficient 3D geometric representations in both explicit and implicit manners based on predicted depth information. Specifically, we utilize the predicted depth to learn voxel occupancy and optimize the voxelized 3D feature volume explicitly through the proposed voxel occupancy attention. To further enhance 3D awareness, the feature volume is integrated with an implicit 3D representation, the truncated signed distance function (TSDF). Without requiring supervision from 3D signals, we significantly improve the model's comprehension of 3D geometry by leveraging intermediate 3D representations and achieve end-to-end training. Our approach surpasses the performance of state-of-the-art image-based methods on both single- and multi-view benchmark datasets across diverse environments, achieving a 9.3 mAP@0.5 improvement on the SUN RGB-D dataset, a 3.3 mAP@0.5 improvement on the ScanNetV2 dataset, and a 0.19 AP3D@0.7 improvement on the KITTI dataset. The project page is available at: https://cindy0725.github.io/3DGeoDet/.
Authors: Ligao Deng, Yupeng Deng, Yu Meng, Jingbo Chen, Zhihao Xi, Diyou Liu, Qifeng Chu
Abstract: Road networks are crucial for mapping, autonomous driving, and disaster response. While manual annotation is costly, deep learning offers efficient extraction. Current methods include postprocessing (prone to errors), global parallel (fast but misses nodes), and local iterative (accurate but slow). We propose GLD-Road, a two-stage model combining global efficiency and local precision. First, it detects road nodes and connects them via a Connect Module. Then, it iteratively refines broken roads using local searches, drastically reducing computation. Experiments show GLD-Road outperforms state-of-the-art methods, improving APLS by 1.9% (City-Scale) and 0.67% (SpaceNet3). It also reduces retrieval time by 40% vs. Sat2Graph (global) and 92% vs. RNGDet++ (local). The experimental results are available at https://github.com/ucas-dlg/GLD-Road.
Authors: Zhaoyang Wei, Chenhui Qiang, Bowen Jiang, Xumeng Han, Xuehui Yu, Zhenjun Han
Abstract: Chain-of-Thought (CoT) reasoning has emerged as a powerful approach to enhance the structured, multi-step decision-making capabilities of Multi-Modal Large Models (MLLMs), is particularly crucial for autonomous driving with adverse weather conditions and complex traffic environments. However, existing benchmarks have largely overlooked the need for rigorous evaluation of CoT processes in these specific and challenging scenarios. To address this critical gap, we introduce AD^2-Bench, the first Chain-of-Thought benchmark specifically designed for autonomous driving with adverse weather and complex scenes. AD^2-Bench is meticulously constructed to fulfill three key criteria: comprehensive data coverage across diverse adverse environments, fine-grained annotations that support multi-step reasoning, and a dedicated evaluation framework tailored for assessing CoT performance. The core contribution of AD^2-Bench is its extensive collection of over 5.4k high-quality, manually annotated CoT instances. Each intermediate reasoning step in these annotations is treated as an atomic unit with explicit ground truth, enabling unprecedented fine-grained analysis of MLLMs' inferential processes under text-level, point-level, and region-level visual prompts. Our comprehensive evaluation of state-of-the-art MLLMs on AD^2-Bench reveals accuracy below 60%, highlighting the benchmark's difficulty and the need to advance robust, interpretable end-to-end autonomous driving systems. AD^2-Bench thus provides a standardized evaluation platform, driving research forward by improving MLLMs' reasoning in autonomous driving, making it an invaluable resource.
Authors: Qijing Li, Jingxiang Sun, Liang An, Zhaoqi Su, Hongwen Zhang, Yebin Liu
Abstract: Holistic 3D scene understanding, which jointly models geometry, appearance, and semantics, is crucial for applications like augmented reality and robotic interaction. Existing feed-forward 3D scene understanding methods (e.g., LSM) are limited to extracting language-based semantics from scenes, failing to achieve holistic scene comprehension. Additionally, they suffer from low-quality geometry reconstruction and noisy artifacts. In contrast, per-scene optimization methods rely on dense input views, which reduces practicality and increases complexity during deployment. In this paper, we propose SemanticSplat, a feed-forward semantic-aware 3D reconstruction method, which unifies 3D Gaussians with latent semantic attributes for joint geometry-appearance-semantics modeling. To predict the semantic anisotropic Gaussians, SemanticSplat fuses diverse feature fields (e.g., LSeg, SAM) with a cost volume representation that stores cross-view feature similarities, enhancing coherent and accurate scene comprehension. Leveraging a two-stage distillation framework, SemanticSplat reconstructs a holistic multi-modal semantic feature field from sparse-view images. Experiments demonstrate the effectiveness of our method for 3D scene understanding tasks like promptable and open-vocabulary segmentation. Video results are available at https://semanticsplat.github.io.
Authors: Mingxiao Li, Mang Ning, Marie-Francine Moens
Abstract: Text-to-image generation models have made significant progress in producing high-quality images from textual descriptions, yet they continue to struggle with maintaining subject consistency across multiple images, a fundamental requirement for visual storytelling. Existing methods attempt to address this by either fine-tuning models on large-scale story visualization datasets, which is resource-intensive, or by using training-free techniques that share information across generations, which still yield limited success. In this paper, we introduce a novel training-free sampling strategy called Zigzag Sampling with Asymmetric Prompts and Visual Sharing to enhance subject consistency in visual story generation. Our approach proposes a zigzag sampling mechanism that alternates between asymmetric prompting to retain subject characteristics, while a visual sharing module transfers visual cues across generated images to %further enforce consistency. Experimental results, based on both quantitative metrics and qualitative evaluations, demonstrate that our method significantly outperforms previous approaches in generating coherent and consistent visual stories. The code is available at https://github.com/Mingxiao-Li/Asymmetry-Zigzag-StoryDiffusion.
URLs: https://github.com/Mingxiao-Li/Asymmetry-Zigzag-StoryDiffusion.
Authors: Giacomo Rosin, Muhammad Rameez Ur Rahman, Sebastiano Vascon
Abstract: Human trajectory forecasting is crucial in applications such as autonomous driving, robotics and surveillance. Accurate forecasting requires models to consider various factors, including social interactions, multi-modal predictions, pedestrian intention and environmental context. While existing methods account for these factors, they often overlook the impact of the environment, which leads to collisions with obstacles. This paper introduces ECAM (Environmental Collision Avoidance Module), a contrastive learning-based module to enhance collision avoidance ability with the environment. The proposed module can be integrated into existing trajectory forecasting models, improving their ability to generate collision-free predictions. We evaluate our method on the ETH/UCY dataset and quantitatively and qualitatively demonstrate its collision avoidance capabilities. Our experiments show that state-of-the-art methods significantly reduce (-40/50%) the collision rate when integrated with the proposed module. The code is available at https://github.com/CVML-CFU/ECAM.
Authors: Yanzhao Shi, Xiaodan Zhang, Junzhong Ji, Haoning Jiang, Chengxin Zheng, Yinong Wang, Liangqiong Qu
Abstract: Automated 3D CT diagnosis empowers clinicians to make timely, evidence-based decisions by enhancing diagnostic accuracy and workflow efficiency. While multimodal large language models (MLLMs) exhibit promising performance in visual-language understanding, existing methods mainly focus on 2D medical images, which fundamentally limits their ability to capture complex 3D anatomical structures. This limitation often leads to misinterpretation of subtle pathologies and causes diagnostic hallucinations. In this paper, we present Hybrid Spatial Encoding Network (HSENet), a framework that exploits enriched 3D medical visual cues by effective visual perception and projection for accurate and robust vision-language understanding. Specifically, HSENet employs dual-3D vision encoders to perceive both global volumetric contexts and fine-grained anatomical details, which are pre-trained by dual-stage alignment with diagnostic reports. Furthermore, we propose Spatial Packer, an efficient multimodal projector that condenses high-resolution 3D spatial regions into a compact set of informative visual tokens via centroid-based compression. By assigning spatial packers with dual-3D vision encoders, HSENet can seamlessly perceive and transfer hybrid visual representations to LLM's semantic space, facilitating accurate diagnostic text generation. Experimental results demonstrate that our method achieves state-of-the-art performance in 3D language-visual retrieval (39.85% of R@100, +5.96% gain), 3D medical report generation (24.01% of BLEU-4, +8.01% gain), and 3D visual question answering (73.60% of Major Class Accuracy, +1.99% gain), confirming its effectiveness. Our code is available at https://github.com/YanzhaoShi/HSENet.
Authors: Dongxu Liu, Yuang Peng, Haomiao Tang, Yuwei Chen, Chunrui Han, Zheng Ge, Daxin Jiang, Mingxue Liao
Abstract: Autoencoders empower state-of-the-art image and video generative models by compressing pixels into a latent space through visual tokenization. Although recent advances have alleviated the performance degradation of autoencoders under high compression ratios, addressing the training instability caused by GAN remains an open challenge. While improving spatial compression, we also aim to minimize the latent space dimensionality, enabling more efficient and compact representations. To tackle these challenges, we focus on improving the decoder's expressiveness. Concretely, we propose DGAE, which employs a diffusion model to guide the decoder in recovering informative signals that are not fully decoded from the latent representation. With this design, DGAE effectively mitigates the performance degradation under high spatial compression rates. At the same time, DGAE achieves state-of-the-art performance with a 2x smaller latent space. When integrated with Diffusion Models, DGAE demonstrates competitive performance on image generation for ImageNet-1K and shows that this compact latent representation facilitates faster convergence of the diffusion model.
Authors: Kunyu Peng, Junchao Huang, Xiangsheng Huang, Di Wen, Junwei Zheng, Yufan Chen, Kailun Yang, Jiamin Wu, Chongqing Hao, Rainer Stiefelhagen
Abstract: Action segmentation is a core challenge in high-level video understanding, aiming to partition untrimmed videos into segments and assign each a label from a predefined action set. Existing methods primarily address single-person activities with fixed action sequences, overlooking multi-person scenarios. In this work, we pioneer textual reference-guided human action segmentation in multi-person settings, where a textual description specifies the target person for segmentation. We introduce the first dataset for Referring Human Action Segmentation, i.e., RHAS133, built from 133 movies and annotated with 137 fine-grained actions with 33h video data, together with textual descriptions for this new task. Benchmarking existing action recognition methods on RHAS133 using VLM-based feature extractors reveals limited performance and poor aggregation of visual cues for the target person. To address this, we propose a holistic-partial aware Fourier-conditioned diffusion framework, i.e., HopaDIFF, leveraging a novel cross-input gate attentional xLSTM to enhance holistic-partial long-range reasoning and a novel Fourier condition to introduce more fine-grained control to improve the action segmentation generation. HopaDIFF achieves state-of-the-art results on RHAS133 in diverse evaluation settings. The code is available at https://github.com/KPeng9510/HopaDIFF.git.
Authors: Haowen Wang, Xiaoping Yuan, Zhao Jin, Zhen Zhao, Zhengping Che, Yousong Xue, Jin Tian, Yakun Huang, Jian Tang
Abstract: Articulated objects are ubiquitous in everyday life, and accurate 3D representations of their geometry and motion are critical for numerous applications. However, in the absence of human annotation, existing approaches still struggle to build a unified representation for objects that contain multiple movable parts. We introduce DeGSS, a unified framework that encodes articulated objects as deformable 3D Gaussian fields, embedding geometry, appearance, and motion in one compact representation. Each interaction state is modeled as a smooth deformation of a shared field, and the resulting deformation trajectories guide a progressive coarse-to-fine part segmentation that identifies distinct rigid components, all in an unsupervised manner. The refined field provides a spatially continuous, fully decoupled description of every part, supporting part-level reconstruction and precise modeling of their kinematic relationships. To evaluate generalization and realism, we enlarge the synthetic PartNet-Mobility benchmark and release RS-Art, a real-to-sim dataset that pairs RGB captures with accurately reverse-engineered 3D models. Extensive experiments demonstrate that our method outperforms existing methods in both accuracy and stability.
Authors: Maik Dannecker, Vasiliki Sideri-Lampretsa, Sophie Starck, Angeline Mihailov, Mathieu Milh, Nadine Girard, Guillaume Auzias, Daniel Rueckert
Abstract: Magnetic resonance imaging of fetal and neonatal brains reveals rapid neurodevelopment marked by substantial anatomical changes unfolding within days. Studying this critical stage of the developing human brain, therefore, requires accurate brain models-referred to as atlases-of high spatial and temporal resolution. To meet these demands, established traditional atlases and recently proposed deep learning-based methods rely on large and comprehensive datasets. This poses a major challenge for studying brains in the presence of pathologies for which data remains scarce. We address this limitation with CINeMA (Conditional Implicit Neural Multi-Modal Atlas), a novel framework for creating high-resolution, spatio-temporal, multimodal brain atlases, suitable for low-data settings. Unlike established methods, CINeMA operates in latent space, avoiding compute-intensive image registration and reducing atlas construction times from days to minutes. Furthermore, it enables flexible conditioning on anatomical features including GA, birth age, and pathologies like ventriculomegaly (VM) and agenesis of the corpus callosum (ACC). CINeMA supports downstream tasks such as tissue segmentation and age prediction whereas its generative properties enable synthetic data creation and anatomically informed data augmentation. Surpassing state-of-the-art methods in accuracy, efficiency, and versatility, CINeMA represents a powerful tool for advancing brain research. We release the code and atlases at https://github.com/m-dannecker/CINeMA.
Authors: Bin Zhu, Hailong Yin, Jingjing Chen, Yu-Gang Jiang
Abstract: Recent advances in reasoning-centric models promise improved robustness through mechanisms such as chain-of-thought prompting and test-time scaling. However, their ability to withstand misleading user input remains underexplored. In this paper, we conduct a systematic evaluation of three state-of-the-art reasoning models, i.e., OpenAI's o4-mini, Claude-3.7-Sonnet and Gemini-2.5-Flash, across three multimodal benchmarks: MMMU, MathVista, and CharXiv. Our evaluation reveals significant accuracy drops (25-29% on average) following gaslighting negation prompts, indicating that even top-tier reasoning models struggle to preserve correct answers under manipulative user feedback. Built upon the insights of the evaluation and to further probe this vulnerability, we introduce GaslightingBench-R, a new diagnostic benchmark specifically designed to evaluate reasoning models' susceptibility to defend their belief under gaslighting negation prompt. Constructed by filtering and curating 1,025 challenging samples from the existing benchmarks, GaslightingBench-R induces even more dramatic failures, with accuracy drops exceeding 53% on average. Our findings reveal fundamental limitations in the robustness of reasoning models, highlighting the gap between step-by-step reasoning and belief persistence.
Authors: Imanol Miranda, Ander Salaberria, Eneko Agirre, Gorka Azkune
Abstract: Dual encoder Vision-Language Models (VLM) such as CLIP are widely used for image-text retrieval tasks. However, those models struggle with compositionality, showing a bag-of-words-like behavior that limits their retrieval performance. Many different training approaches have been proposed to improve the vision-language compositionality capabilities of those models. In comparison, inference-time techniques have received little attention. In this paper, we propose to add simple structure at inference, where, given an image and a caption: i) we divide the image into different smaller crops, ii) we extract text segments, capturing objects, attributes and relations, iii) using a VLM, we find the image crops that better align with text segments obtaining matches, and iv) we compute the final image-text similarity aggregating the individual similarities of the matches. Based on various popular dual encoder VLMs, we evaluate our approach in controlled and natural datasets for VL compositionality. We find that our approach consistently improves the performance of evaluated VLMs without any training, which shows the potential of inference-time techniques. The results are especially good for attribute-object binding as shown in the controlled dataset. As a result of an extensive analysis: i) we show that processing image crops is actually essential for the observed gains in performance, and ii) we identify specific areas to further improve inference-time approaches.
Authors: Changwei Wu, Yifei Chen, Yuxin Du, Jinying Zong, Jie Dong, Mingxuan Liu, Yong Peng, Jin Fan, Feiwei Qin, Changmiao Wang
Abstract: Early diagnosis of Alzheimer's Disease (AD), especially at the mild cognitive impairment (MCI) stage, is vital yet hindered by subjective assessments and the high cost of multimodal imaging modalities. Although deep learning methods offer automated alternatives, their energy inefficiency and computational demands limit real-world deployment, particularly in resource-constrained settings. As a brain-inspired paradigm, spiking neural networks (SNNs) are inherently well-suited for modeling the sparse, event-driven patterns of neural degeneration in AD, offering a promising foundation for interpretable and low-power medical diagnostics. However, existing SNNs often suffer from weak expressiveness and unstable training, which restrict their effectiveness in complex medical tasks. To address these limitations, we propose FasterSNN, a hybrid neural architecture that integrates biologically inspired LIF neurons with region-adaptive convolution and multi-scale spiking attention. This design enables sparse, efficient processing of 3D MRI while preserving diagnostic accuracy. Experiments on benchmark datasets demonstrate that FasterSNN achieves competitive performance with substantially improved efficiency and stability, supporting its potential for practical AD screening. Our source code is available at https://github.com/wuchangw/FasterSNN.
Authors: Mattia Nardon, Mikel Mujika Agirre, Ander Gonz\'alez Tom\'e, Daniel Sedano Algarabel, Josep Rueda Collell, Ana Paola Caro, Andrea Caraffa, Fabio Poiesi, Paul Ian Chippendale, Davide Boscaini
Abstract: Accurate 6D pose estimation of complex objects in 3D environments is essential for effective robotic manipulation. Yet, existing benchmarks fall short in evaluating 6D pose estimation methods under realistic industrial conditions, as most datasets focus on household objects in domestic settings, while the few available industrial datasets are limited to artificial setups with objects placed on tables. To bridge this gap, we introduce CHIP, the first dataset designed for 6D pose estimation of chairs manipulated by a robotic arm in a real-world industrial environment. CHIP includes seven distinct chairs captured using three different RGBD sensing technologies and presents unique challenges, such as distractor objects with fine-grained differences and severe occlusions caused by the robotic arm and human operators. CHIP comprises 77,811 RGBD images annotated with ground-truth 6D poses automatically derived from the robot's kinematics, averaging 11,115 annotations per chair. We benchmark CHIP using three zero-shot 6D pose estimation methods, assessing performance across different sensor types, localization priors, and occlusion levels. Results show substantial room for improvement, highlighting the unique challenges posed by the dataset. CHIP will be publicly released.
Authors: Xulin Ma, Jiankai Tang, Zhang Jiang, Songqin Cheng, Yuanchun Shi, Dong LI, Xin Liu, Daniel McDuff, Xiaojing Liu, Yuntao Wang
Abstract: Remote photoplethysmography (rPPG) enables non-contact, continuous monitoring of physiological signals and offers a practical alternative to traditional health sensing methods. Although rPPG is promising for daily health monitoring, its application in long-term personal care scenarios, such as mirror-facing routines in high-altitude environments, remains challenging due to ambient lighting variations, frequent occlusions from hand movements, and dynamic facial postures. To address these challenges, we present LADH (Long-term Altitude Daily Health), the first long-term rPPG dataset containing 240 synchronized RGB and infrared (IR) facial videos from 21 participants across five common personal care scenarios, along with ground-truth PPG, respiration, and blood oxygen signals. Our experiments demonstrate that combining RGB and IR video inputs improves the accuracy and robustness of non-contact physiological monitoring, achieving a mean absolute error (MAE) of 4.99 BPM in heart rate estimation. Furthermore, we find that multi-task learning enhances performance across multiple physiological indicators simultaneously. Dataset and code are open at https://github.com/McJackTang/FusionVitals.
Authors: Ye Zhang, Yu Zhou, Yifeng Wang, Jun Xiao, Ziyue Wang, Yongbing Zhang, Jianxu Chen
Abstract: Cell instance segmentation is critical to analyzing biomedical images, yet accurately distinguishing tightly touching cells remains a persistent challenge. Existing instance segmentation frameworks, including detection-based, contour-based, and distance mapping-based approaches, have made significant progress, but balancing model performance with computational efficiency remains an open problem. In this paper, we propose a novel cell instance segmentation method inspired by the four-color theorem. By conceptualizing cells as countries and tissues as oceans, we introduce a four-color encoding scheme that ensures adjacent instances receive distinct labels. This reformulation transforms instance segmentation into a constrained semantic segmentation problem with only four predicted classes, substantially simplifying the instance differentiation process. To solve the training instability caused by the non-uniqueness of four-color encoding, we design an asymptotic training strategy and encoding transformation method. Extensive experiments on various modes demonstrate our approach achieves state-of-the-art performance. The code is available at https://github.com/zhangye-zoe/FCIS.
Authors: Chuang Ma, Shaokai Zhao, Dongdong Zhou, Yu Pei, Zhiguo Luo, Liang Xie, Ye Yan, Erwei Yin
Abstract: Micro-expression recognition (MER), a critical subfield of affective computing, presents greater challenges than macro-expression recognition due to its brief duration and low intensity. While incorporating prior knowledge has been shown to enhance MER performance, existing methods predominantly rely on simplistic, singular sources of prior knowledge, failing to fully exploit multi-source information. This paper introduces the Multi-Prior Fusion Network (MPFNet), leveraging a progressive training strategy to optimize MER tasks. We propose two complementary encoders: the Generic Feature Encoder (GFE) and the Advanced Feature Encoder (AFE), both based on Inflated 3D ConvNets (I3D) with Coordinate Attention (CA) mechanisms, to improve the model's ability to capture spatiotemporal and channel-specific features. Inspired by developmental psychology, we present two variants of MPFNet--MPFNet-P and MPFNet-C--corresponding to two fundamental modes of infant cognitive development: parallel and hierarchical processing. These variants enable the evaluation of different strategies for integrating prior knowledge. Extensive experiments demonstrate that MPFNet significantly improves MER accuracy while maintaining balanced performance across categories, achieving accuracies of 0.811, 0.924, and 0.857 on the SMIC, CASME II, and SAMM datasets, respectively. To the best of our knowledge, our approach achieves state-of-the-art performance on the SMIC and SAMM datasets.
Authors: Yuting Li, Lai Wei, Kaipeng Zheng, Jingyuan Huang, Linghe Kong, Lichao Sun, Weiran Huang
Abstract: Despite the rapid progress of multimodal large language models (MLLMs), they have largely overlooked the importance of visual processing. In a simple yet revealing experiment, we interestingly find that language-only models, when provided with image captions, can achieve comparable or even better performance than MLLMs that consume raw visual inputs. This suggests that current MLLMs may generate accurate visual descriptions but fail to effectively integrate them during reasoning. Motivated by this, we propose a simple visual perturbation framework that enhances perceptual robustness without requiring algorithmic modifications or additional training data. Our approach introduces three targeted perturbations: distractor concatenation, dominance-preserving mixup, and random rotation, that can be easily integrated into existing post-training pipelines including SFT, DPO, and GRPO. Through extensive experiments across multiple datasets, we demonstrate consistent improvements in mathematical reasoning performance, with gains comparable to those achieved through algorithmic changes. Additionally, we achieve competitive performance among open-source 7B RL-tuned models by training Qwen2.5-VL-7B with visual perturbation. Through comprehensive ablation studies, we analyze the effectiveness of different perturbation strategies, revealing that each perturbation type contributes uniquely to different aspects of visual reasoning. Our findings highlight the critical role of visual perturbation in multimodal mathematical reasoning: better reasoning begins with better seeing. Our code is available at https://github.com/YutingLi0606/Vision-Matters.
Authors: Qin Zhou, Zhiyang Zhang, Jinglong Wang, Xiaobin Li, Jing Zhang, Qian Yu, Lu Sheng, Dong Xu
Abstract: Diffusion models excel at image generation. Recent studies have shown that these models not only generate high-quality images but also encode text-image alignment information through attention maps or loss functions. This information is valuable for various downstream tasks, including segmentation, text-guided image editing, and compositional image generation. However, current methods heavily rely on the assumption of perfect text-image alignment in diffusion models, which is not the case. In this paper, we propose using zero-shot referring image segmentation as a proxy task to evaluate the pixel-level image and class-level text alignment of popular diffusion models. We conduct an in-depth analysis of pixel-text misalignment in diffusion models from the perspective of training data bias. We find that misalignment occurs in images with small sized, occluded, or rare object classes. Therefore, we propose ELBO-T2IAlign, a simple yet effective method to calibrate pixel-text alignment in diffusion models based on the evidence lower bound (ELBO) of likelihood. Our method is training-free and generic, eliminating the need to identify the specific cause of misalignment and works well across various diffusion model architectures. Extensive experiments on commonly used benchmark datasets on image segmentation and generation have verified the effectiveness of our proposed calibration approach.
Authors: Yangrui Zhu, Junhua Bao, Yipan Wei, Yapeng Li, Bo Du
Abstract: Existing multimodal methods typically assume that different modalities share the same category set. However, in real-world applications, the category distributions in multimodal data exhibit inconsistencies, which can hinder the model's ability to effectively utilize cross-modal information for recognizing all categories. In this work, we propose the practical setting termed Multi-Modal Heterogeneous Category-set Learning (MMHCL), where models are trained in heterogeneous category sets of multi-modal data and aim to recognize complete classes set of all modalities during test. To effectively address this task, we propose a Class Similarity-based Cross-modal Fusion model (CSCF). Specifically, CSCF aligns modality-specific features to a shared semantic space to enable knowledge transfer between seen and unseen classes. It then selects the most discriminative modality for decision fusion through uncertainty estimation. Finally, it integrates cross-modal information based on class similarity, where the auxiliary modality refines the prediction of the dominant one. Experimental results show that our method significantly outperforms existing state-of-the-art (SOTA) approaches on multiple benchmark datasets, effectively addressing the MMHCL task.
Authors: Xiangkai Zhang, Xiang Zhou, Mao Chen, Yuchen Lu, Xu Yang, Zhiyong Liu
Abstract: Absolute localization, aiming to determine an agent's location with respect to a global reference, is crucial for unmanned aerial vehicles (UAVs) in various applications, but it becomes challenging when global navigation satellite system (GNSS) signals are unavailable. Vision-based absolute localization methods, which locate the current view of the UAV in a reference satellite map to estimate its position, have become popular in GNSS-denied scenarios. However, existing methods mostly rely on traditional and low-level image matching, suffering from difficulties due to significant differences introduced by cross-source discrepancies and temporal variations. To overcome these limitations, in this paper, we introduce a hierarchical cross-source image matching method designed for UAV absolute localization, which integrates a semantic-aware and structure-constrained coarse matching module with a lightweight fine-grained matching module. Specifically, in the coarse matching module, semantic features derived from a vision foundation model first establish region-level correspondences under semantic and structural constraints. Then, the fine-grained matching module is applied to extract fine features and establish pixel-level correspondences. Building upon this, a UAV absolute visual localization pipeline is constructed without any reliance on relative localization techniques, mainly by employing an image retrieval module before the proposed hierarchical image matching modules. Experimental evaluations on public benchmark datasets and a newly introduced CS-UAV dataset demonstrate superior accuracy and robustness of the proposed method under various challenging conditions, confirming its effectiveness.
Authors: Anton Razzhigaev, Matvey Mikhalchuk, Klim Kireev, Igor Udovichenko, Andrey Kuznetsov, Aleksandr Petiushko
Abstract: Reconstructing facial images from black-box recognition models poses a significant privacy threat. While many methods require access to embeddings, we address the more challenging scenario of model inversion using only similarity scores. This paper introduces DarkerBB, a novel approach that reconstructs color faces by performing zero-order optimization within a PCA-derived eigenface space. Despite this highly limited information, experiments on LFW, AgeDB-30, and CFP-FP benchmarks demonstrate that DarkerBB achieves state-of-the-art verification accuracies in the similarity-only setting, with competitive query efficiency.
Authors: Nicola Farronato, Florian Scheidegger, Mattia Rigotti, Cristiano Malossi, Michele Magno, Haotong Qin
Abstract: The Segment Anything Model 2 (SAM2) has gained significant attention as a foundational approach for promptable image and video segmentation. However, its expensive computational and memory consumption poses a severe challenge for its application in resource-constrained scenarios. In this paper, we propose an accurate low-bit quantization method for efficient SAM2, termed Q-SAM2. To address the performance degradation caused by the singularities in weight and activation distributions during quantization, Q-SAM2 introduces two novel technical contributions. We first introduce a linear layer calibration method for low-bit initialization of SAM2, which minimizes the Frobenius norm over a small image batch to reposition weight distributions for improved quantization. We then propose a Quantization-Aware Training (QAT) pipeline that applies clipping to suppress outliers and allows the network to adapt to quantization thresholds during training. Our comprehensive experiments demonstrate that Q-SAM2 allows for highly accurate inference while substantially improving efficiency. Both quantitative and visual results show that our Q-SAM2 surpasses existing state-of-the-art general quantization schemes, especially for ultra-low 2-bit quantization. While designed for quantization-aware training, our proposed calibration technique also proves effective in post-training quantization, achieving up to a 66% mIoU accuracy improvement over non-calibrated models.
Authors: Andrea Caraffa, Davide Boscaini, Fabio Poiesi
Abstract: Estimating the 6D pose of objects from RGBD data is a fundamental problem in computer vision, with applications in robotics and augmented reality. A key challenge is achieving generalization to novel objects that were not seen during training. Most existing approaches address this by scaling up training on synthetic data tailored to the task, a process that demands substantial computational resources. But is task-specific training really necessary for accurate and efficient 6D pose estimation of novel objects? To answer No!, we introduce FreeZeV2, the second generation of FreeZe: a training-free method that achieves strong generalization to unseen objects by leveraging geometric and vision foundation models pre-trained on unrelated data. FreeZeV2 improves both accuracy and efficiency over FreeZe through three key contributions: (i) a sparse feature extraction strategy that reduces inference-time computation without sacrificing accuracy; (ii) a feature-aware scoring mechanism that improves both pose selection during RANSAC-based 3D registration and the final ranking of pose candidates; and (iii) a modular design that supports ensembles of instance segmentation models, increasing robustness to segmentation masks errors. We evaluate FreeZeV2 on the seven core datasets of the BOP Benchmark, where it establishes a new state-of-the-art in 6D pose estimation of unseen objects. When using the same segmentation masks, FreeZeV2 achieves a remarkable 8x speedup over FreeZe while also improving accuracy by 5%. When using ensembles of segmentation models, FreeZeV2 gains an additional 8% in accuracy while still running 2.5x faster than FreeZe. FreeZeV2 was awarded Best Overall Method at the BOP Challenge 2024.
Authors: Xiandong Zou, Ruihao Xia, Hongsong Wang, Pan Zhou
Abstract: While text-to-3D generation has attracted growing interest, existing methods often struggle to produce 3D assets that align well with human preferences. Current preference alignment techniques for 3D content typically rely on hardly-collected preference-paired multi-view 2D images to train 2D reward models, when then guide 3D generation -- leading to geometric artifacts due to their inherent 2D bias. To address these limitations, we construct 3D-MeshPref, the first large-scale unpaired 3D preference dataset, featuring diverse 3D meshes annotated by a large language model and refined by human evaluators. We then develop RewardCS, the first reward model trained directly on unpaired 3D-MeshPref data using a novel Cauchy-Schwarz divergence objective, enabling effective learning of human-aligned 3D geometric preferences without requiring paired comparisons. Building on this, we propose DreamCS, a unified framework that integrates RewardCS into text-to-3D pipelines -- enhancing both implicit and explicit 3D generation with human preference feedback. Extensive experiments show DreamCS outperforms prior methods, producing 3D assets that are both geometrically faithful and human-preferred. Code and models will be released publicly.
Authors: Chuang Ma, Yu Pei, Jianhang Zhang, Shaokai Zhao, Bowen Ji, Liang Xie, Ye Yan, Erwei Yin
Abstract: Micro-expressions (MEs) are subtle, fleeting nonverbal cues that reveal an individual's genuine emotional state. Their analysis has attracted considerable interest due to its promising applications in fields such as healthcare, criminal investigation, and human-computer interaction. However, existing ME research is limited to single visual modality, overlooking the rich emotional information conveyed by other physiological modalities, resulting in ME recognition and spotting performance far below practical application needs. Therefore, exploring the cross-modal association mechanism between ME visual features and physiological signals (PS), and developing a multimodal fusion framework, represents a pivotal step toward advancing ME analysis. This study introduces a novel ME dataset, MMME, which, for the first time, enables synchronized collection of facial action signals (MEs), central nervous system signals (EEG), and peripheral PS (PPG, RSP, SKT, EDA, and ECG). By overcoming the constraints of existing ME corpora, MMME comprises 634 MEs, 2,841 macro-expressions (MaEs), and 2,890 trials of synchronized multimodal PS, establishing a robust foundation for investigating ME neural mechanisms and conducting multimodal fusion-based analyses. Extensive experiments validate the dataset's reliability and provide benchmarks for ME analysis, demonstrating that integrating MEs with PS significantly enhances recognition and spotting performance. To the best of our knowledge, MMME is the most comprehensive ME dataset to date in terms of modality diversity. It provides critical data support for exploring the neural mechanisms of MEs and uncovering the visual-physiological synergistic effects, driving a paradigm shift in ME research from single-modality visual analysis to multimodal fusion. The dataset will be publicly available upon acceptance of this paper.
Authors: Junli Deng, Ping Shi, Qipei Li, Jinyang Guo
Abstract: Reconstructing intricate, ever-changing environments remains a central ambition in computer vision, yet existing solutions often crumble before the complexity of real-world dynamics. We present DynaSplat, an approach that extends Gaussian Splatting to dynamic scenes by integrating dynamic-static separation and hierarchical motion modeling. First, we classify scene elements as static or dynamic through a novel fusion of deformation offset statistics and 2D motion flow consistency, refining our spatial representation to focus precisely where motion matters. We then introduce a hierarchical motion modeling strategy that captures both coarse global transformations and fine-grained local movements, enabling accurate handling of intricate, non-rigid motions. Finally, we integrate physically-based opacity estimation to ensure visually coherent reconstructions, even under challenging occlusions and perspective shifts. Extensive experiments on challenging datasets reveal that DynaSplat not only surpasses state-of-the-art alternatives in accuracy and realism but also provides a more intuitive, compact, and efficient route to dynamic scene reconstruction.
Authors: Chen Gao, Liankai Jin, Xingyu Peng, Jiazhao Zhang, Yue Deng, Annan Li, He Wang, Si Liu
Abstract: Embodied navigation stands as a foundation pillar within the broader pursuit of embodied AI. However, previous navigation research is divided into different tasks/capabilities, e.g., ObjNav, ImgNav and VLN, where they differ in task objectives and modalities, making datasets and methods are designed individually. In this work, we take steps toward generalist navigation agents, which can follow free-form instructions that include arbitrary compounds of multi-modal and multi-capability. To achieve this, we propose a large-scale benchmark and corresponding method, termed OctoNav-Bench and OctoNav-R1. Specifically, OctoNav-Bench features continuous environments and is constructed via a designed annotation pipeline. We thoroughly craft instruction-trajectory pairs, where instructions are diverse in free-form with arbitrary modality and capability. Also, we construct a Think-Before-Action (TBA-CoT) dataset within OctoNav-Bench to provide the thinking process behind actions. For OctoNav-R1, we build it upon MLLMs and adapt it to a VLA-type model, which can produce low-level actions solely based on 2D visual observations. Moreover, we design a Hybrid Training Paradigm (HTP) that consists of three stages, i.e., Action-/TBA-SFT, Nav-GPRO, and Online RL stages. Each stage contains specifically designed learning policies and rewards. Importantly, for TBA-SFT and Nav-GRPO designs, we are inspired by the OpenAI-o1 and DeepSeek-R1, which show impressive reasoning ability via thinking-before-answer. Thus, we aim to investigate how to achieve thinking-before-action in the embodied navigation field, to improve model's reasoning ability toward generalists. Specifically, we propose TBA-SFT to utilize the TBA-CoT dataset to fine-tune the model as a cold-start phrase and then leverage Nav-GPRO to improve its thinking ability. Finally, OctoNav-R1 shows superior performance compared with previous methods.
Authors: Panagiotis Kaliosis, John Pavlopoulos
Abstract: Handwritten text recognition aims to convert visual input into machine-readable text, and it remains challenging due to the evolving and context-dependent nature of handwriting. Character sets change over time, and character frequency distributions shift across historical periods or regions, often causing models trained on broad, heterogeneous corpora to underperform on specific subsets. To tackle this, we propose a novel loss function that incorporates the Wasserstein distance between the character frequency distribution of the predicted text and a target distribution empirically derived from training data. By penalizing divergence from expected distributions, our approach enhances both accuracy and robustness under temporal and contextual intra-dataset shifts. Furthermore, we demonstrate that character distribution alignment can also improve existing models at inference time without requiring retraining by integrating it as a scoring function in a guided decoding scheme. Experimental results across multiple datasets and architectures confirm the effectiveness of our method in boosting generalization and performance. We open source our code at https://github.com/pkaliosis/fada.
Authors: Florian Bordes, Quentin Garrido, Justine T Kao, Adina Williams, Michael Rabbat, Emmanuel Dupoux
Abstract: We present IntPhys 2, a video benchmark designed to evaluate the intuitive physics understanding of deep learning models. Building on the original IntPhys benchmark, IntPhys 2 focuses on four core principles related to macroscopic objects: Permanence, Immutability, Spatio-Temporal Continuity, and Solidity. These conditions are inspired by research into intuitive physical understanding emerging during early childhood. IntPhys 2 offers a comprehensive suite of tests, based on the violation of expectation framework, that challenge models to differentiate between possible and impossible events within controlled and diverse virtual environments. Alongside the benchmark, we provide performance evaluations of several state-of-the-art models. Our findings indicate that while these models demonstrate basic visual understanding, they face significant challenges in grasping intuitive physics across the four principles in complex scenes, with most models performing at chance levels (50%), in stark contrast to human performance, which achieves near-perfect accuracy. This underscores the gap between current models and human-like intuitive physics understanding, highlighting the need for advancements in model architectures and training methodologies.
Authors: Siyu Chen, Ting Han, Chengzheng Fu, Changshe Zhang, Chaolei Wang, Jinhe Su, Guorong Cai, Meiliu Wu
Abstract: Open-Vocabulary semantic segmentation (OVSS) and domain generalization in semantic segmentation (DGSS) highlight a subtle complementarity that motivates Open-Vocabulary Domain-Generalized Semantic Segmentation (OV-DGSS). OV-DGSS aims to generate pixel-level masks for unseen categories while maintaining robustness across unseen domains, a critical capability for real-world scenarios such as autonomous driving in adverse conditions. We introduce Vireo, a novel single-stage framework for OV-DGSS that unifies the strengths of OVSS and DGSS for the first time. Vireo builds upon the frozen Visual Foundation Models (VFMs) and incorporates scene geometry via Depth VFMs to extract domain-invariant structural features. To bridge the gap between visual and textual modalities under domain shift, we propose three key components: (1) GeoText Prompts, which align geometric features with language cues and progressively refine VFM encoder representations; (2) Coarse Mask Prior Embedding (CMPE) for enhancing gradient flow for faster convergence and stronger textual influence; and (3) the Domain-Open-Vocabulary Vector Embedding Head (DOV-VEH), which fuses refined structural and semantic features for robust prediction. Comprehensive evaluation on these components demonstrates the effectiveness of our designs. Our proposed Vireo achieves the state-of-the-art performance and surpasses existing methods by a large margin in both domain generalization and open-vocabulary recognition, offering a unified and scalable solution for robust visual understanding in diverse and dynamic environments. Code is available at https://github.com/anonymouse-9c53tp182bvz/Vireo.
Authors: Seonho Lee, Jiho Choi, Inha Kang, Jiwook Kim, Junsung Park, Hyunjung Shim
Abstract: Vision-Language Models (VLMs) have shown remarkable performance on diverse visual and linguistic tasks, yet they remain fundamentally limited in their understanding of 3D spatial structures. We propose Geometric Distillation, a lightweight, annotation-free fine-tuning framework that injects human-inspired geometric cues into pretrained VLMs without modifying their architecture. By distilling (1) sparse correspondences, (2) relative depth relations, and (3) dense cost volumes from off-the-shelf 3D foundation models (e.g., MASt3R, VGGT), our method shapes representations to be geometry-aware while remaining compatible with natural image-text inputs. Through extensive evaluations on 3D vision-language reasoning and 3D perception benchmarks, our method consistently outperforms prior approaches, achieving improved 3D spatial reasoning with significantly lower computational cost. Our work demonstrates a scalable and efficient path to bridge 2D-trained VLMs with 3D understanding, opening up wider use in spatially grounded multimodal tasks.
Authors: Haoru Wang, Kai Ye, Yangyan Li, Wenzheng Chen, Baoquan Chen
Abstract: We consider the problem of generalizable novel view synthesis (NVS), which aims to generate photorealistic novel views from sparse or even unposed 2D images without per-scene optimization. This task remains fundamentally challenging, as it requires inferring 3D structure from incomplete and ambiguous 2D observations. Early approaches typically rely on strong 3D knowledge, including architectural 3D inductive biases (e.g., embedding explicit 3D representations, such as NeRF or 3DGS, into network design) and ground-truth camera poses for both input and target views. While recent efforts have sought to reduce the 3D inductive bias or the dependence on known camera poses of input views, critical questions regarding the role of 3D knowledge and the necessity of circumventing its use remain under-explored. In this work, we conduct a systematic analysis on the 3D knowledge and uncover a critical trend: the performance of methods that requires less 3D knowledge accelerates more as data scales, eventually achieving performance on par with their 3D knowledge-driven counterparts, which highlights the increasing importance of reducing dependence on 3D knowledge in the era of large-scale data. Motivated by and following this trend, we propose a novel NVS framework that minimizes 3D inductive bias and pose dependence for both input and target views. By eliminating this 3D knowledge, our method fully leverages data scaling and learns implicit 3D awareness directly from sparse 2D images, without any 3D inductive bias or pose annotation during training. Extensive experiments demonstrate that our model generates photorealistic and 3D-consistent novel views, achieving even comparable performance with methods that rely on posed inputs, thereby validating the feasibility and effectiveness of our data-centric paradigm. Project page: https://pku-vcl-geometry.github.io/Less3Depend/ .
Authors: Athinoulla Konstantinou, Georgios Leontidis, Mamatha Thota, Aiden Durrant
Abstract: Learning self-supervised representations that are invariant and equivariant to transformations is crucial for advancing beyond traditional visual classification tasks. However, many methods rely on predictor architectures to encode equivariance, despite evidence that architectural choices, such as capsule networks, inherently excel at learning interpretable pose-aware representations. To explore this, we introduce EquiCaps (Equivariant Capsule Network), a capsule-based approach to pose-aware self-supervision that eliminates the need for a specialised predictor for enforcing equivariance. Instead, we leverage the intrinsic pose-awareness capabilities of capsules to improve performance in pose estimation tasks. To further challenge our assumptions, we increase task complexity via multi-geometric transformations to enable a more thorough evaluation of invariance and equivariance by introducing 3DIEBench-T, an extension of a 3D object-rendering benchmark dataset. Empirical results demonstrate that EquiCaps outperforms prior state-of-the-art equivariant methods on rotation prediction, achieving a supervised-level $R^2$ of 0.78 on the 3DIEBench rotation prediction benchmark and improving upon SIE and CapsIE by 0.05 and 0.04 $R^2$, respectively. Moreover, in contrast to non-capsule-based equivariant approaches, EquiCaps maintains robust equivariant performance under combined geometric transformations, underscoring its generalisation capabilities and the promise of predictor-free capsule architectures.
Authors: Tao Liu, Zhenchao Cui
Abstract: Tiny object detection (TOD) reveals a fundamental flaw in feature pyramid networks: high-level features (P5-P6) frequently receive zero positive anchors under standard label assignment protocols, leaving their semantic representations untrained due to exclusion from loss computation. This creates dual deficiencies: (1) Stranded high-level features become semantic dead-ends without gradient updates, while (2) low-level features lack essential semantic context for robust classification. We propose E-FPN-BS that systematically converts wasted high-level semantics into low-level feature enhancements. To address these issues, we propose E-FPN-BS, a novel architecture integrating multi-scale feature enhancement and adaptive optimization. First, our Context Enhancement Module(CEM) employs dual-branch processing to align and compress high-level features for effective global-local fusion. Second, the Foreground-Background Separation Module (FBSM) generates spatial gating masks that dynamically amplify discriminative regions. To address gradient imbalance across object scales, we further propose a Dynamic Gradient-Balanced Loss (DCLoss) that automatically modulates loss contributions via scale-aware gradient equilibrium. Extensive experiments across multiple benchmark datasets demonstrate the outstanding performance and generalization ability of our approach.
Authors: Tilemachos Aravanis (School of Electrical & Computer Engineering, National Technical University of Athens, Greece), Panagiotis Filntisis (Robotics Institute, Athena Research Center, Maroussi, Greece, HERON - Center of Excellence in Robotics, Athens, Greece), Petros Maragos (School of Electrical & Computer Engineering, National Technical University of Athens, Greece, Robotics Institute, Athena Research Center, Maroussi, Greece, HERON - Center of Excellence in Robotics, Athens, Greece), George Retsinas (Robotics Institute, Athena Research Center, Maroussi, Greece, HERON - Center of Excellence in Robotics, Athens, Greece)
Abstract: Generating images in a consistent reference visual style remains a challenging computer vision task. State-of-the-art methods aiming for style-consistent generation struggle to effectively separate semantic content from stylistic elements, leading to content leakage from the image provided as a reference to the targets. To address this challenge, we propose Only-Style: a method designed to mitigate content leakage in a semantically coherent manner while preserving stylistic consistency. Only-Style works by localizing content leakage during inference, allowing the adaptive tuning of a parameter that controls the style alignment process, specifically within the image patches containing the subject in the reference image. This adaptive process best balances stylistic consistency with leakage elimination. Moreover, the localization of content leakage can function as a standalone component, given a reference-target image pair, allowing the adaptive tuning of any method-specific parameter that provides control over the impact of the stylistic reference. In addition, we propose a novel evaluation framework to quantify the success of style-consistent generations in avoiding undesired content leakage. Our approach demonstrates a significant improvement over state-of-the-art methods through extensive evaluation across diverse instances, consistently achieving robust stylistic consistency without undesired content leakage.
Authors: He Zhang, Chentao Song, Hongwen Zhang, Tao Yu
Abstract: We introduce MetricHMR (Metric Human Mesh Recovery), an approach for metric human mesh recovery with accurate global translation from monocular images. In contrast to existing HMR methods that suffer from severe scale and depth ambiguity, MetricHMR is able to produce geometrically reasonable body shape and global translation in the reconstruction results. To this end, we first systematically analyze previous HMR methods on camera models to emphasize the critical role of the standard perspective projection model in enabling metric-scale HMR. We then validate the acceptable ambiguity range of metric HMR under the standard perspective projection model. Finally, we contribute a novel approach that introduces a ray map based on the standard perspective projection to jointly encode bounding-box information, camera parameters, and geometric cues for End2End metric HMR without any additional metric-regularization modules. Extensive experiments demonstrate that our method achieves state-of-the-art performance, even compared with sequential HMR methods, in metric pose, shape, and global translation estimation across both indoor and in-the-wild scenarios.
Authors: Jianhan Qi, Yuheng Jia, Hui Liu, Junhui Hou
Abstract: Hyperspectral image (HSI) clustering assigns similar pixels to the same class without any annotations, which is an important yet challenging task. For large-scale HSIs, most methods rely on superpixel segmentation and perform superpixel-level clustering based on graph neural networks (GNNs). However, existing GNNs cannot fully exploit the spectral information of the input HSI, and the inaccurate superpixel topological graph may lead to the confusion of different class semantics during information aggregation. To address these challenges, we first propose a structural-spectral graph convolutional operator (SSGCO) tailored for graph-structured HSI superpixels to improve their representation quality through the co-extraction of spatial and spectral features. Second, we propose an evidence-guided adaptive edge learning (EGAEL) module that adaptively predicts and refines edge weights in the superpixel topological graph. We integrate the proposed method into a contrastive learning framework to achieve clustering, where representation learning and clustering are simultaneously conducted. Experiments demonstrate that the proposed method improves clustering accuracy by 2.61%, 6.06%, 4.96% and 3.15% over the best compared methods on four HSI datasets. Our code is available at https://github.com/jhqi/SSGCO-EGAEL.
Authors: Marco Federici, Riccardo Del Chiaro, Boris van Breugel, Paul Whatmough, Markus Nagel
Abstract: Diffusion models represent the cutting edge in image generation, but their high memory and computational demands hinder deployment on resource-constrained devices. Post-Training Quantization (PTQ) offers a promising solution by reducing the bitwidth of matrix operations. However, standard PTQ methods struggle with outliers, and achieving higher compression often requires transforming model weights and activations before quantization. In this work, we propose HadaNorm, a novel linear transformation that extends existing approaches and effectively mitigates outliers by normalizing activations feature channels before applying Hadamard transformations, enabling more aggressive activation quantization. We demonstrate that HadaNorm consistently reduces quantization error across the various components of transformer blocks, achieving superior efficiency-performance trade-offs when compared to state-of-the-art methods.
Authors: Jiangyong Huang, Xiaojian Ma, Xiongkun Linghu, Yue Fan, Junchao He, Wenxin Tan, Qing Li, Song-Chun Zhu, Yixin Chen, Baoxiong Jia, Siyuan Huang
Abstract: Developing 3D-VL generalists capable of understanding 3D scenes and following natural language instructions to perform a wide range of tasks has been a long-standing goal in the 3D-VL community. Despite recent progress, 3D-VL models still lag behind their 2D counterparts in capability and robustness, falling short of the generalist standard. A key obstacle to developing 3D-VL generalists lies in data scalability, hindered by the lack of an efficient scene representation. We propose LEO-VL, a 3D-VL model built upon condensed feature grid (CFG), an efficient scene representation that bridges 2D perception and 3D spatial structure while significantly reducing token overhead. This efficiency unlocks large-scale training towards 3D-VL generalist, for which we curate over 700k high-quality 3D-VL data spanning four domains of real-world indoor scenes and five tasks such as captioning and dialogue. LEO-VL achieves state-of-the-art performance on a variety of 3D QA benchmarks, including SQA3D, MSQA, and Beacon3D. Ablation studies confirm the efficiency of our representation, the importance of task and scene diversity, and the validity of our data curation principle. Furthermore, we introduce SceneDPO, a novel post-training objective that enhances the robustness of 3D-VL models. We hope our findings contribute to the advancement of scalable and robust 3D-VL generalists.
Authors: Aaron Foss, Chloe Evans, Sasha Mitts, Koustuv Sinha, Ammar Rizvi, Justine T. Kao
Abstract: We introduce CausalVQA, a benchmark dataset for video question answering (VQA) composed of question-answer pairs that probe models' understanding of causality in the physical world. Existing VQA benchmarks either tend to focus on surface perceptual understanding of real-world videos, or on narrow physical reasoning questions created using simulation environments. CausalVQA fills an important gap by presenting challenging questions that are grounded in real-world scenarios, while focusing on models' ability to predict the likely outcomes of different actions and events through five question types: counterfactual, hypothetical, anticipation, planning and descriptive. We designed quality control mechanisms that prevent models from exploiting trivial shortcuts, requiring models to base their answers on deep visual understanding instead of linguistic cues. We find that current frontier multimodal models fall substantially below human performance on the benchmark, especially on anticipation and hypothetical questions. This highlights a challenge for current systems to leverage spatial-temporal reasoning, understanding of physical principles, and comprehension of possible alternatives to make accurate predictions in real-world settings.
Authors: Ziyi Wang, Yanran Zhang, Jie Zhou, Jiwen Lu
Abstract: The scale diversity of point cloud data presents significant challenges in developing unified representation learning techniques for 3D vision. Currently, there are few unified 3D models, and no existing pre-training method is equally effective for both object- and scene-level point clouds. In this paper, we introduce UniPre3D, the first unified pre-training method that can be seamlessly applied to point clouds of any scale and 3D models of any architecture. Our approach predicts Gaussian primitives as the pre-training task and employs differentiable Gaussian splatting to render images, enabling precise pixel-level supervision and end-to-end optimization. To further regulate the complexity of the pre-training task and direct the model's focus toward geometric structures, we integrate 2D features from pre-trained image models to incorporate well-established texture knowledge. We validate the universal effectiveness of our proposed method through extensive experiments across a variety of object- and scene-level tasks, using diverse point cloud models as backbones. Code is available at https://github.com/wangzy22/UniPre3D.
Authors: Benjamin Reichman, Constantin Patsch, Jack Truxal, Atishay Jain, Larry Heck
Abstract: In outside knowledge visual question answering (OK-VQA), the model must identify relevant visual information within an image and incorporate external knowledge to accurately respond to a question. Extending this task to a visually grounded dialogue setting based on videos, a conversational model must both recognize pertinent visual details over time and answer questions where the required information is not necessarily present in the visual information. Moreover, the context of the overall conversation must be considered for the subsequent dialogue. To explore this task, we introduce a dataset comprised of $2,017$ videos with $5,986$ human-annotated dialogues consisting of $40,954$ interleaved dialogue turns. While the dialogue context is visually grounded in specific video segments, the questions further require external knowledge that is not visually present. Thus, the model not only has to identify relevant video parts but also leverage external knowledge to converse within the dialogue. We further provide several baselines evaluated on our dataset and show future challenges associated with this task. The dataset is made publicly available here: https://github.com/c-patsch/OKCV.
Authors: Ziyi Wang, Yongming Rao, Shuofeng Sun, Xinrun Liu, Yi Wei, Xumin Yu, Zuyan Liu, Yanbo Wang, Hongmin Liu, Jie Zhou, Jiwen Lu
Abstract: Recently, we have witnessed the great success of the generalist model in natural language processing. The generalist model is a general framework trained with massive data and is able to process various downstream tasks simultaneously. Encouraged by their impressive performance, an increasing number of researchers are venturing into the realm of applying these models to computer vision tasks. However, the inputs and outputs of vision tasks are more diverse, and it is difficult to summarize them as a unified representation. In this paper, we provide a comprehensive overview of the vision generalist models, delving into their characteristics and capabilities within the field. First, we review the background, including the datasets, tasks, and benchmarks. Then, we dig into the design of frameworks that have been proposed in existing research, while also introducing the techniques employed to enhance their performance. To better help the researchers comprehend the area, we take a brief excursion into related domains, shedding light on their interconnections and potential synergies. To conclude, we provide some real-world application scenarios, undertake a thorough examination of the persistent challenges, and offer insights into possible directions for future research endeavors.
Authors: Sushant Gautam, Michael A. Riegler, P{\aa}l Halvorsen
Abstract: Medical Visual Question Answering (MedVQA) is a promising field for developing clinical decision support systems, yet progress is often limited by the available datasets, which can lack clinical complexity and visual diversity. To address these gaps, we introduce Kvasir-VQA-x1, a new, large-scale dataset for gastrointestinal (GI) endoscopy. Our work significantly expands upon the original Kvasir-VQA by incorporating 159,549 new question-answer pairs that are designed to test deeper clinical reasoning. We developed a systematic method using large language models to generate these questions, which are stratified by complexity to better assess a model's inference capabilities. To ensure our dataset prepares models for real-world clinical scenarios, we have also introduced a variety of visual augmentations that mimic common imaging artifacts. The dataset is structured to support two main evaluation tracks: one for standard VQA performance and another to test model robustness against these visual perturbations. By providing a more challenging and clinically relevant benchmark, Kvasir-VQA-x1 aims to accelerate the development of more reliable and effective multimodal AI systems for use in clinical settings. The dataset is fully accessible and adheres to FAIR data principles, making it a valuable resource for the wider research community. Code and data: https://github.com/Simula/Kvasir-VQA-x1 and https://huggingface.co/datasets/SimulaMet/Kvasir-VQA-x1
URLs: https://github.com/Simula/Kvasir-VQA-x1, https://huggingface.co/datasets/SimulaMet/Kvasir-VQA-x1
Authors: Junfei Wu, Jian Guan, Kaituo Feng, Qiang Liu, Shu Wu, Liang Wang, Wei Wu, Tieniu Tan
Abstract: As textual reasoning with large language models (LLMs) has advanced significantly, there has been growing interest in enhancing the multimodal reasoning capabilities of large vision-language models (LVLMs). However, existing methods primarily approach multimodal reasoning in a straightforward, text-centric manner, where both reasoning and answer derivation are conducted purely through text, with the only difference being the presence of multimodal input. As a result, these methods often encounter fundamental limitations in spatial reasoning tasks that demand precise geometric understanding and continuous spatial tracking-capabilities that humans achieve through mental visualization and manipulation. To address the limitations, we propose drawing to reason in space, a novel paradigm that enables LVLMs to reason through elementary drawing operations in the visual space. By equipping models with basic drawing operations, including annotating bounding boxes and drawing auxiliary lines, we empower them to express and analyze spatial relationships through direct visual manipulation, meanwhile avoiding the performance ceiling imposed by specialized perception tools in previous tool-integrated reasoning approaches. To cultivate this capability, we develop a three-stage training framework: cold-start training with synthetic data to establish basic drawing abilities, reflective rejection sampling to enhance self-reflection behaviors, and reinforcement learning to directly optimize for target rewards. Extensive experiments demonstrate that our model, named VILASR, consistently outperforms existing methods across diverse spatial reasoning benchmarks, involving maze navigation, static spatial reasoning, video-based reasoning, and multi-view-based reasoning tasks, with an average improvement of 18.4%.
Authors: Jeripothula Prudviraj, Vikram Jamwal
Abstract: Creating a stroke-by-stroke evolution process of a visual artwork tries to bridge the emotional and educational gap between the finished static artwork and its creation process. Recent stroke-based painting systems focus on capturing stroke details by predicting and iteratively refining stroke parameters to maximize the similarity between the input image and the rendered output. However, these methods often struggle to produce stroke compositions that align with artistic principles and intent. To address this, we explore an image-to-painting method that (i) facilitates semantic guidance for brush strokes in targeted regions, (ii) computes the brush stroke parameters, and (iii) establishes a sequence among segments and strokes to sequentially render the final painting. Experimental results on various input image types, such as face images, paintings, and photographic images, show that our method aligns with a region-based painting strategy while rendering a painting with high fidelity and superior stroke quality.
Authors: Jiaxiang Tang, Ruijie Lu, Zhaoshuo Li, Zekun Hao, Xuan Li, Fangyin Wei, Shuran Song, Gang Zeng, Ming-Yu Liu, Tsung-Yi Lin
Abstract: Recent progress in 3D object generation has greatly improved both the quality and efficiency. However, most existing methods generate a single mesh with all parts fused together, which limits the ability to edit or manipulate individual parts. A key challenge is that different objects may have a varying number of parts. To address this, we propose a new end-to-end framework for part-level 3D object generation. Given a single input image, our method generates high-quality 3D objects with an arbitrary number of complete and semantically meaningful parts. We introduce a dual volume packing strategy that organizes all parts into two complementary volumes, allowing for the creation of complete and interleaved parts that assemble into the final object. Experiments show that our model achieves better quality, diversity, and generalization than previous image-based part-level generation methods.
Authors: Jiazhi Yang, Kashyap Chitta, Shenyuan Gao, Long Chen, Yuqian Shao, Xiaosong Jia, Hongyang Li, Andreas Geiger, Xiangyu Yue, Li Chen
Abstract: How can we reliably simulate future driving scenarios under a wide range of ego driving behaviors? Recent driving world models, developed exclusively on real-world driving data composed mainly of safe expert trajectories, struggle to follow hazardous or non-expert behaviors, which are rare in such data. This limitation restricts their applicability to tasks such as policy evaluation. In this work, we address this challenge by enriching real-world human demonstrations with diverse non-expert data collected from a driving simulator (e.g., CARLA), and building a controllable world model trained on this heterogeneous corpus. Starting with a video generator featuring a diffusion transformer architecture, we devise several strategies to effectively integrate conditioning signals and improve prediction controllability and fidelity. The resulting model, ReSim, enables Reliable Simulation of diverse open-world driving scenarios under various actions, including hazardous non-expert ones. To close the gap between high-fidelity simulation and applications that require reward signals to judge different actions, we introduce a Video2Reward module that estimates a reward from ReSim's simulated future. Our ReSim paradigm achieves up to 44% higher visual fidelity, improves controllability for both expert and non-expert actions by over 50%, and boosts planning and policy selection performance on NAVSIM by 2% and 25%, respectively.
Authors: Zijie Wu, Chaohui Yu, Fan Wang, Xiang Bai
Abstract: Recent advances in 4D content generation have attracted increasing attention, yet creating high-quality animated 3D models remains challenging due to the complexity of modeling spatio-temporal distributions and the scarcity of 4D training data. In this paper, we present AnimateAnyMesh, the first feed-forward framework that enables efficient text-driven animation of arbitrary 3D meshes. Our approach leverages a novel DyMeshVAE architecture that effectively compresses and reconstructs dynamic mesh sequences by disentangling spatial and temporal features while preserving local topological structures. To enable high-quality text-conditional generation, we employ a Rectified Flow-based training strategy in the compressed latent space. Additionally, we contribute the DyMesh Dataset, containing over 4M diverse dynamic mesh sequences with text annotations. Experimental results demonstrate that our method generates semantically accurate and temporally coherent mesh animations in a few seconds, significantly outperforming existing approaches in both quality and efficiency. Our work marks a substantial step forward in making 4D content creation more accessible and practical. All the data, code, and models will be open-released.
Authors: Zhenzhi Wang, Jiaqi Yang, Jianwen Jiang, Chao Liang, Gaojie Lin, Zerong Zheng, Ceyuan Yang, Dahua Lin
Abstract: End-to-end human animation with rich multi-modal conditions, e.g., text, image and audio has achieved remarkable advancements in recent years. However, most existing methods could only animate a single subject and inject conditions in a global manner, ignoring scenarios that multiple concepts could appears in the same video with rich human-human interactions and human-object interactions. Such global assumption prevents precise and per-identity control of multiple concepts including humans and objects, therefore hinders applications. In this work, we discard the single-entity assumption and introduce a novel framework that enforces strong, region-specific binding of conditions from modalities to each identity's spatiotemporal footprint. Given reference images of multiple concepts, our method could automatically infer layout information by leveraging a mask predictor to match appearance cues between the denoised video and each reference appearance. Furthermore, we inject local audio condition into its corresponding region to ensure layout-aligned modality matching in a iterative manner. This design enables the high-quality generation of controllable multi-concept human-centric videos. Empirical results and ablation studies validate the effectiveness of our explicit layout control for multi-modal conditions compared to implicit counterparts and other existing methods.
Authors: Benno Krojer, Mojtaba Komeili, Candace Ross, Quentin Garrido, Koustuv Sinha, Nicolas Ballas, Mahmoud Assran
Abstract: Existing benchmarks for assessing the spatio-temporal understanding and reasoning abilities of video language models are susceptible to score inflation due to the presence of shortcut solutions based on superficial visual or textual cues. This paper mitigates the challenges in accurately assessing model performance by introducing the Minimal Video Pairs (MVP) benchmark, a simple shortcut-aware video QA benchmark for assessing the physical understanding of video language models. The benchmark is comprised of 55K high-quality multiple-choice video QA examples focusing on physical world understanding. Examples are curated from nine video data sources, spanning first-person egocentric and exocentric videos, robotic interaction data, and cognitive science intuitive physics benchmarks. To mitigate shortcut solutions that rely on superficial visual or textual cues and biases, each sample in MVP has a minimal-change pair -- a visually similar video accompanied by an identical question but an opposing answer. To answer a question correctly, a model must provide correct answers for both examples in the minimal-change pair; as such, models that solely rely on visual or textual biases would achieve below random performance. Human performance on MVP is 92.9\%, while the best open-source state-of-the-art video-language model achieves 40.2\% compared to random performance at 25\%.
Authors: Ron Yosef, Moran Yanuka, Yonatan Bitton, Dani Lischinski
Abstract: Text-guided image editing, fueled by recent advancements in generative AI, is becoming increasingly widespread. This trend highlights the need for a comprehensive framework to verify text-guided edits and assess their quality. To address this need, we introduce EditInspector, a novel benchmark for evaluation of text-guided image edits, based on human annotations collected using an extensive template for edit verification. We leverage EditInspector to evaluate the performance of state-of-the-art (SoTA) vision and language models in assessing edits across various dimensions, including accuracy, artifact detection, visual quality, seamless integration with the image scene, adherence to common sense, and the ability to describe edit-induced changes. Our findings indicate that current models struggle to evaluate edits comprehensively and frequently hallucinate when describing the changes. To address these challenges, we propose two novel methods that outperform SoTA models in both artifact detection and difference caption generation.
Authors: Yiming Dou, Wonseok Oh, Yuqing Luo, Antonio Loquercio, Andrew Owens
Abstract: We study the problem of making 3D scene reconstructions interactive by asking the following question: can we predict the sounds of human hands physically interacting with a scene? First, we record a video of a human manipulating objects within a 3D scene using their hands. We then use these action-sound pairs to train a rectified flow model to map 3D hand trajectories to their corresponding audio. At test time, a user can query the model for other actions, parameterized as sequences of hand poses, to estimate their corresponding sounds. In our experiments, we find that our generated sounds accurately convey material properties and actions, and that they are often indistinguishable to human observers from real sounds. Project page: https://www.yimingdou.com/hearing_hands/
Authors: Jaewon Min, Jin Hyeon Kim, Paul Hyunbin Cho, Jaeeun Lee, Jihye Park, Minkyu Park, Sangpil Kim, Hyunhee Park, Seungryong Kim
Abstract: Image restoration aims to recover degraded images. However, existing diffusion-based restoration methods, despite great success in natural image restoration, often struggle to faithfully reconstruct textual regions in degraded images. Those methods frequently generate plausible but incorrect text-like patterns, a phenomenon we refer to as text-image hallucination. In this paper, we introduce Text-Aware Image Restoration (TAIR), a novel restoration task that requires the simultaneous recovery of visual contents and textual fidelity. To tackle this task, we present SA-Text, a large-scale benchmark of 100K high-quality scene images densely annotated with diverse and complex text instances. Furthermore, we propose a multi-task diffusion framework, called TeReDiff, that integrates internal features from diffusion models into a text-spotting module, enabling both components to benefit from joint training. This allows for the extraction of rich text representations, which are utilized as prompts in subsequent denoising steps. Extensive experiments demonstrate that our approach consistently outperforms state-of-the-art restoration methods, achieving significant gains in text recognition accuracy. See our project page: https://cvlab-kaist.github.io/TAIR/
Authors: Yuanpeng Tu, Hao Luo, Xi Chen, Xiang Bai, Fan Wang, Hengshuang Zhao
Abstract: We introduce PlayerOne, the first egocentric realistic world simulator, facilitating immersive and unrestricted exploration within vividly dynamic environments. Given an egocentric scene image from the user, PlayerOne can accurately construct the corresponding world and generate egocentric videos that are strictly aligned with the real scene human motion of the user captured by an exocentric camera. PlayerOne is trained in a coarse-to-fine pipeline that first performs pretraining on large-scale egocentric text-video pairs for coarse-level egocentric understanding, followed by finetuning on synchronous motion-video data extracted from egocentric-exocentric video datasets with our automatic construction pipeline. Besides, considering the varying importance of different components, we design a part-disentangled motion injection scheme, enabling precise control of part-level movements. In addition, we devise a joint reconstruction framework that progressively models both the 4D scene and video frames, ensuring scene consistency in the long-form video generation. Experimental results demonstrate its great generalization ability in precise control of varying human movements and worldconsistent modeling of diverse scenarios. It marks the first endeavor into egocentric real-world simulation and can pave the way for the community to delve into fresh frontiers of world modeling and its diverse applications.
Authors: Shayan Shekarforoush, David B. Lindell, Marcus A. Brubaker, David J. Fleet
Abstract: Cryo-EM is a transformational paradigm in molecular biology where computational methods are used to infer 3D molecular structure at atomic resolution from extremely noisy 2D electron microscope images. At the forefront of research is how to model the structure when the imaged particles exhibit non-rigid conformational flexibility and compositional variation where parts are sometimes missing. We introduce a novel 3D reconstruction framework with a hierarchical Gaussian mixture model, inspired in part by Gaussian Splatting for 4D scene reconstruction. In particular, the structure of the model is grounded in an initial process that infers a part-based segmentation of the particle, providing essential inductive bias in order to handle both conformational and compositional variability. The framework, called CryoSPIRE, is shown to reveal biologically meaningful structures on complex experimental datasets, and establishes a new state-of-the-art on CryoBench, a benchmark for cryo-EM heterogeneity methods.
Authors: Abigail Copiaco, Christian Ritz, Yassine Himeur, Valsamma Eapen, Ammar Albanna, Wathiq Mansoor
Abstract: The prevalence of Autism Spectrum Disorder (ASD) has surged rapidly over the past decade, posing significant challenges in communication, behavior, and focus for affected individuals. Current diagnostic techniques, though effective, are time-intensive, leading to high social and economic costs. This work introduces an AI-powered assistive technology designed to streamline ASD diagnosis and management, enhancing convenience for individuals with ASD and efficiency for caregivers and therapists. The system integrates transfer learning with image transforms derived from eye gaze variables to diagnose ASD. This facilitates and opens opportunities for in-home periodical diagnosis, reducing stress for individuals and caregivers, while also preserving user privacy through the use of image transforms. The accessibility of the proposed method also offers opportunities for improved communication between guardians and therapists, ensuring regular updates on progress and evolving support needs. Overall, the approach proposed in this work ensures timely, accessible diagnosis while protecting the subjects' privacy, improving outcomes for individuals with ASD.
Authors: Sahaj Raj Malla
Abstract: Handwritten digit recognition in regional scripts, such as Devanagari, is crucial for multilingual document digitization, educational tools, and the preservation of cultural heritage. The script's complex structure and limited annotated datasets pose significant challenges to conventional models. This paper introduces the first hybrid quantum-classical architecture for Devanagari handwritten digit recognition, combining a convolutional neural network (CNN) for spatial feature extraction with a 10-qubit variational quantum circuit (VQC) for quantum-enhanced classification. Trained and evaluated on the Devanagari Handwritten Character Dataset (DHCD), the proposed model achieves a state-of-the-art test accuracy for quantum implementation of 99.80% and a test loss of 0.2893, with an average per-class F1-score of 0.9980. Compared to equivalent classical CNNs, our model demonstrates superior accuracy with significantly fewer parameters and enhanced robustness. By leveraging quantum principles such as superposition and entanglement, this work establishes a novel benchmark for regional script recognition, highlighting the promise of quantum machine learning (QML) in real-world, low-resource language settings.
Authors: Elly Akhoundi, Hung Yu Ling, Anup Anand Deshmukh, Judith Butepage
Abstract: Motion in-betweening is a crucial tool for animators, enabling intricate control over pose-level details in each keyframe. Recent machine learning solutions for motion in-betweening rely on complex models, incorporating skeleton-aware architectures or requiring multiple modules and training steps. In this work, we introduce a simple yet effective Transformer-based framework, employing a single Transformer encoder to synthesize realistic motions for motion in-betweening tasks. We find that data modeling choices play a significant role in improving in-betweening performance. Among others, we show that increasing data volume can yield equivalent or improved motion transitions, that the choice of pose representation is vital for achieving high-quality results, and that incorporating velocity input features enhances animation performance. These findings challenge the assumption that model complexity is the primary determinant of animation quality and provide insights into a more data-centric approach to motion interpolation. Additional videos and supplementary material are available at https://silk-paper.github.io.
Authors: Vivien van Veldhuizen, Vanessa Botha, Chunyao Lu, Melis Erdal Cesur, Kevin Groot Lipman, Edwin D. de Jong, Hugo Horlings, Cl\'arisa Sanchez, Cees Snoek, Ritse Mann, Eric Marcus, Jonas Teuwen
Abstract: Foundation models (FMs) are changing the way medical images are analyzed by learning from large collections of unlabeled data. Instead of relying on manually annotated examples, FMs are pre-trained to learn general-purpose visual features that can later be adapted to specific clinical tasks with little additional supervision. In this review, we examine how FMs are being developed and applied in pathology, radiology, and ophthalmology, drawing on evidence from over 150 studies. We explain the core components of FM pipelines, including model architectures, self-supervised learning methods, and strategies for downstream adaptation. We also review how FMs are being used in each imaging domain and compare design choices across applications. Finally, we discuss key challenges and open questions to guide future research.
Authors: Yangjie Cui, Boyang Gao, Yiwei Zhang, Xin Dong, Jinwu Xiang, Daochun Li, Zhan Tu
Abstract: Previous studies on event camera sensing have demonstrated certain detection performance using dense event representations. However, the accumulated noise in such dense representations has received insufficient attention, which degrades the representation quality and increases the likelihood of missed detections. To address this challenge, we propose the Wavelet Denoising-enhanced DEtection TRansformer, i.e., WD-DETR network, for event cameras. In particular, a dense event representation is presented first, which enables real-time reconstruction of events as tensors. Then, a wavelet transform method is designed to filter noise in the event representations. Such a method is integrated into the backbone for feature extraction. The extracted features are subsequently fed into a transformer-based network for object prediction. To further reduce inference time, we incorporate the Dynamic Reorganization Convolution Block (DRCB) as a fusion module within the hybrid encoder. The proposed method has been evaluated on three event-based object detection datasets, i.e., DSEC, Gen1, and 1Mpx. The results demonstrate that WD-DETR outperforms tested state-of-the-art methods. Additionally, we implement our approach on a common onboard computer for robots, the NVIDIA Jetson Orin NX, achieving a high frame rate of approximately 35 FPS using TensorRT FP16, which is exceptionally well-suited for real-time perception of onboard robotic systems.
Authors: Haonan Zhang, Guoyan Lao, Yuyao Zhang, Hongjiang Wei
Abstract: Quantitative magnetic resonance imaging (qMRI) provides tissue-specific parameters vital for clinical diagnosis. Although simultaneous multi-parametric qMRI (MP-qMRI) technologies enhance imaging efficiency, robustly reconstructing qMRI from highly undersampled, high-dimensional measurements remains a significant challenge. This difficulty arises primarily because current reconstruction methods that rely solely on a single prior or physics-informed model to solve the highly ill-posed inverse problem, which often leads to suboptimal results. To overcome this limitation, we propose LoREIN, a novel unsupervised and dual-prior-integrated framework for accelerated 3D MP-qMRI reconstruction. Technically, LoREIN incorporates both low-rank prior and continuity prior via low-rank representation (LRR) and implicit neural representation (INR), respectively, to enhance reconstruction fidelity. The powerful continuous representation of INR enables the estimation of optimal spatial bases within the low-rank subspace, facilitating high-fidelity reconstruction of weighted images. Simultaneously, the predicted multi-contrast weighted images provide essential structural and quantitative guidance, further enhancing the reconstruction accuracy of quantitative parameter maps. Furthermore, our work introduces a zero-shot learning paradigm with broad potential in complex spatiotemporal and high-dimensional image reconstruction tasks, further advancing the field of medical imaging.
Authors: Rajan Das Gupta, Md Imrul Hasan Showmick, Mushfiqur Rahman Abir, Shanjida Akter, Md. Yeasin Rahat, Md. Jakir Hossen
Abstract: Early and accurate detection of brain abnormalities, such as tumors and strokes, is essential for timely intervention and improved patient outcomes. In this study, we present a deep learning-based system capable of identifying both brain tumors and strokes from MRI images, along with their respective stages. We have executed two groundbreaking strategies involving convolutional neural networks, MobileNet V2 and ResNet-50-optimized through transfer learning to classify MRI scans into five diagnostic categories. Our dataset, aggregated and augmented from various publicly available MRI sources, was carefully curated to ensure class balance and image diversity. To enhance model generalization and prevent overfitting, we applied dropout layers and extensive data augmentation. The models achieved strong performance, with training accuracy reaching 93\% and validation accuracy up to 88\%. While ResNet-50 demonstrated slightly better results, Mobile Net V2 remains a promising option for real-time diagnosis in low resource settings due to its lightweight architecture. This research offers a practical AI-driven solution for early brain abnormality detection, with potential for clinical deployment and future enhancement through larger datasets and multi modal inputs.
Authors: Tyler J. Richards, Adam E. Flanders, Errol Colak, Luciano M. Prevedello, Robyn L. Ball, Felipe Kitamura, John Mongan, Maryam Vazirabad, Hui-Ming Lin, Anne Kendell, Thanat Kanthawang, Salita Angkurawaranon, Emre Altinmakas, Hakan Dogan, Paulo Eduardo de Aguiar Kuriki, Arjuna Somasundaram, Christopher Ruston, Deniz Bulja, Naida Spahovic, Jennifer Sommer, Sirui Jiang, Eduardo Moreno Judice de Mattos Farina, Eduardo Caminha Nunes, Michael Brassil, Megan McNamara, Johanna Ortiz, Jacob Peoples, Vinson L. Uytana, Anthony Kam, Venkata N. S. Dola, Daniel Murphy, David Vu, Dataset Contributor Group, Dataset Annotator Group, Competition Data Notebook Group, Jason F. Talbott
Abstract: The Radiological Society of North America (RSNA) Lumbar Degenerative Imaging Spine Classification (LumbarDISC) dataset is the largest publicly available dataset of adult MRI lumbar spine examinations annotated for degenerative changes. The dataset includes 2,697 patients with a total of 8,593 image series from 8 institutions across 6 countries and 5 continents. The dataset is available for free for non-commercial use via Kaggle and RSNA Medical Imaging Resource of AI (MIRA). The dataset was created for the RSNA 2024 Lumbar Spine Degenerative Classification competition where competitors developed deep learning models to grade degenerative changes in the lumbar spine. The degree of spinal canal, subarticular recess, and neural foraminal stenosis was graded at each intervertebral disc level in the lumbar spine. The images were annotated by expert volunteer neuroradiologists and musculoskeletal radiologists from the RSNA, American Society of Neuroradiology, and the American Society of Spine Radiology. This dataset aims to facilitate research and development in machine learning and lumbar spine imaging to lead to improved patient care and clinical efficiency.
Authors: Pranav Guruprasad, Yangyue Wang, Harshvardhan Sikka
Abstract: Recent innovations in multimodal action models represent a promising direction for developing general-purpose agentic systems, combining visual understanding, language comprehension, and action generation. We introduce MultiNet - a novel, fully open-source benchmark and surrounding software ecosystem designed to rigorously evaluate and adapt models across vision, language, and action domains. We establish standardized evaluation protocols for assessing vision-language models (VLMs) and vision-language-action models (VLAs), and provide open source software to download relevant data, models, and evaluations. Additionally, we provide a composite dataset with over 1.3 trillion tokens of image captioning, visual question answering, commonsense reasoning, robotic control, digital game-play, simulated locomotion/manipulation, and many more tasks. The MultiNet benchmark, framework, toolkit, and evaluation harness have been used in downstream research on the limitations of VLA generalization.
Authors: Boyu Jiang, Liang Shi, Zhengzhi Lin, Loren Stowe, Feng Guo
Abstract: The performance of perception systems in autonomous driving systems (ADS) is strongly influenced by object distance, scene dynamics, and environmental conditions such as weather. AI-based perception outputs are inherently stochastic, with variability driven by these external factors, while traditional evaluation metrics remain static and event-independent, failing to capture fluctuations in confidence over time. In this work, we introduce the Perception Characteristics Distance (PCD) -- a novel evaluation metric that quantifies the farthest distance at which an object can be reliably detected, incorporating uncertainty in model outputs. To support this, we present the SensorRainFall dataset, collected on the Virginia Smart Road using a sensor-equipped vehicle (cameras, radar, LiDAR) under controlled daylight-clear and daylight-rain scenarios, with precise ground-truth distances to the target objects. Statistical analysis reveals the presence of change points in the variance of detection confidence score with distance. By averaging the PCD values across a range of detection quality thresholds and probabilistic thresholds, we compute the mean PCD (mPCD), which captures the overall perception characteristics of a system with respect to detection distance. Applying state-of-the-art perception models shows that mPCD captures meaningful reliability differences under varying weather conditions -- differences that static metrics overlook. PCD provides a principled, distribution-aware measure of perception performance, supporting safer and more robust ADS operation, while the SensorRainFall dataset offers a valuable benchmark for evaluation. The SensorRainFall dataset is publicly available at https://www.kaggle.com/datasets/datadrivenwheels/sensorrainfall, and the evaluation code is open-sourced at https://github.com/datadrivenwheels/PCD_Python.
URLs: https://www.kaggle.com/datasets/datadrivenwheels/sensorrainfall,, https://github.com/datadrivenwheels/PCD_Python.
Authors: Yihe Tang, Wenlong Huang, Yingke Wang, Chengshu Li, Roy Yuan, Ruohan Zhang, Jiajun Wu, Li Fei-Fei
Abstract: Understanding fine-grained object affordances is imperative for robots to manipulate objects in unstructured environments given open-ended task instructions. However, existing methods of visual affordance predictions often rely on manually annotated data or conditions only on a predefined set of tasks. We introduce UAD (Unsupervised Affordance Distillation), a method for distilling affordance knowledge from foundation models into a task-conditioned affordance model without any manual annotations. By leveraging the complementary strengths of large vision models and vision-language models, UAD automatically annotates a large-scale dataset with detailed $<$instruction, visual affordance$>$ pairs. Training only a lightweight task-conditioned decoder atop frozen features, UAD exhibits notable generalization to in-the-wild robotic scenes and to various human activities, despite only being trained on rendered objects in simulation. Using affordance provided by UAD as the observation space, we show an imitation learning policy that demonstrates promising generalization to unseen object instances, object categories, and even variations in task instructions after training on as few as 10 demonstrations. Project website: https://unsup-affordance.github.io/
Authors: Inclusion AI, Biao Gong, Cheng Zou, Chuanyang Zheng, Chunluan Zhou, Canxiang Yan, Chunxiang Jin, Chunjie Shen, Dandan Zheng, Fudong Wang, Furong Xu, GuangMing Yao, Jun Zhou, Jingdong Chen, Jianxin Sun, Jiajia Liu, Jianjiang Zhu, Jun Peng, Kaixiang Ji, Kaiyou Song, Kaimeng Ren, Libin Wang, Lixiang Ru, Lele Xie, Longhua Tan, Lyuxin Xue, Lan Wang, Mochen Bai, Ning Gao, Pei Chen, Qingpei Guo, Qinglong Zhang, Qiang Xu, Rui Liu, Ruijie Xiong, Sirui Gao, Tinghao Liu, Taisong Li, Weilong Chai, Xinyu Xiao, Xiaomei Wang, Xiaoxue Chen, Xiao Lu, Xiaoyu Li, Xingning Dong, Xuzheng Yu, Yi Yuan, Yuting Gao, Yunxiao Sun, Yipeng Chen, Yifei Wu, Yongjie Lyu, Ziping Ma, Zipeng Feng, Zhijiang Fang, Zhihao Qiu, Ziyuan Huang, Zhengyu He
Abstract: We propose Ming-Omni, a unified multimodal model capable of processing images, text, audio, and video, while demonstrating strong proficiency in both speech and image generation. Ming-Omni employs dedicated encoders to extract tokens from different modalities, which are then processed by Ling, an MoE architecture equipped with newly proposed modality-specific routers. This design enables a single model to efficiently process and fuse multimodal inputs within a unified framework, thereby facilitating diverse tasks without requiring separate models, task-specific fine-tuning, or structural redesign. Importantly, Ming-Omni extends beyond conventional multimodal models by supporting audio and image generation. This is achieved through the integration of an advanced audio decoder for natural-sounding speech and Ming-Lite-Uni for high-quality image generation, which also allow the model to engage in context-aware chatting, perform text-to-speech conversion, and conduct versatile image editing. Our experimental results showcase Ming-Omni offers a powerful solution for unified perception and generation across all modalities. Notably, our proposed Ming-Omni is the first open-source model we are aware of to match GPT-4o in modality support, and we release all code and model weights to encourage further research and development in the community.
Authors: Yitong Zhang, Jia Li, Liyi Cai, Ge Li
Abstract: Large Vision-Language Models (LVLMs) have achieved impressive progress across various applications but remain vulnerable to malicious queries that exploit the visual modality. Existing alignment approaches typically fail to resist malicious queries while preserving utility on benign ones effectively. To address these challenges, we propose Deep Aligned Visual Safety Prompt (DAVSP), which is built upon two key innovations. First, we introduce the Visual Safety Prompt, which appends a trainable padding region around the input image. It preserves visual features and expands the optimization space. Second, we propose Deep Alignment, a novel approach to train the visual safety prompt through supervision in the model's activation space. It enhances the inherent ability of LVLMs to perceive malicious queries, achieving deeper alignment than prior works. Extensive experiments across five benchmarks on two representative LVLMs demonstrate that DAVSP effectively resists malicious queries while preserving benign input utility. Furthermore, DAVSP exhibits great cross-model generation ability. Ablation studies further reveal that both the Visual Safety Prompt and Deep Alignment are essential components, jointly contributing to its overall effectiveness. The code is publicly available at https://github.com/zhangyitonggg/DAVSP.
Authors: Jiaqi Tang, Yu Xia, Yi-Feng Wu, Yuwei Hu, Yuhui Chen, Qing-Guo Chen, Xiaogang Xu, Xiangyu Wu, Hao Lu, Yanqing Ma, Shiyin Lu, Qifeng Chen
Abstract: The advent of autonomous agents is transforming interactions with Graphical User Interfaces (GUIs) by employing natural language as a powerful intermediary. Despite the predominance of Supervised Fine-Tuning (SFT) methods in current GUI agents for achieving spatial localization, these methods face substantial challenges due to their limited capacity to accurately perceive positional data. Existing strategies, such as reinforcement learning, often fail to assess positional accuracy effectively, thereby restricting their utility. In response, we introduce Location Preference Optimization (LPO), a novel approach that leverages locational data to optimize interaction preferences. LPO uses information entropy to predict interaction positions by focusing on zones rich in information. Besides, it further introduces a dynamic location reward function based on physical distance, reflecting the varying importance of interaction positions. Supported by Group Relative Preference Optimization (GRPO), LPO facilitates an extensive exploration of GUI environments and significantly enhances interaction precision. Comprehensive experiments demonstrate LPO's superior performance, achieving SOTA results across both offline benchmarks and real-world online evaluations. Our code will be made publicly available soon, at https://github.com/AIDC-AI/LPO.
Authors: Guanghu Xie, Zhiduo Jiang, Yonglong Zhang, Yang Liu, Zongwu Xie, Baoshi Cao, Hong Liu
Abstract: Transparent and reflective objects in everyday environments pose significant challenges for depth sensors due to their unique visual properties, such as specular reflections and light transmission. These characteristics often lead to incomplete or inaccurate depth estimation, which severely impacts downstream geometry-based vision tasks, including object recognition, scene reconstruction, and robotic manipulation. To address the issue of missing depth information in transparent and reflective objects, we propose DCIRNet, a novel multimodal depth completion network that effectively integrates RGB images and depth maps to enhance depth estimation quality. Our approach incorporates an innovative multimodal feature fusion module designed to extract complementary information between RGB images and incomplete depth maps. Furthermore, we introduce a multi-stage supervision and depth refinement strategy that progressively improves depth completion and effectively mitigates the issue of blurred object boundaries. We integrate our depth completion model into dexterous grasping frameworks and achieve a $44\%$ improvement in the grasp success rate for transparent and reflective objects. We conduct extensive experiments on public datasets, where DCIRNet demonstrates superior performance. The experimental results validate the effectiveness of our approach and confirm its strong generalization capability across various transparent and reflective objects.
Authors: Shuai Wang, Zhenhua Liu, Jiaheng Wei, Xuanwu Yin, Dong Li, Emad Barsoum
Abstract: We present Athena-PRM, a multimodal process reward model (PRM) designed to evaluate the reward score for each step in solving complex reasoning problems. Developing high-performance PRMs typically demands significant time and financial investment, primarily due to the necessity for step-level annotations of reasoning steps. Conventional automated labeling methods, such as Monte Carlo estimation, often produce noisy labels and incur substantial computational costs. To efficiently generate high-quality process-labeled data, we propose leveraging prediction consistency between weak and strong completers as a criterion for identifying reliable process labels. Remarkably, Athena-PRM demonstrates outstanding effectiveness across various scenarios and benchmarks with just 5,000 samples. Furthermore, we also develop two effective strategies to improve the performance of PRMs: ORM initialization and up-sampling for negative data. We validate our approach in three specific scenarios: verification for test time scaling, direct evaluation of reasoning step correctness, and reward ranked fine-tuning. Our Athena-PRM consistently achieves superior performance across multiple benchmarks and scenarios. Notably, when using Qwen2.5-VL-7B as the policy model, Athena-PRM enhances performance by 10.2 points on WeMath and 7.1 points on MathVista for test time scaling. Furthermore, Athena-PRM sets the state-of-the-art (SoTA) results in VisualProcessBench and outperforms the previous SoTA by 3.9 F1-score, showcasing its robust capability to accurately assess the correctness of the reasoning step. Additionally, utilizing Athena-PRM as the reward model, we develop Athena-7B with reward ranked fine-tuning and outperforms baseline with a significant margin on five benchmarks.
Authors: Fatemeh Mohammadi Amin, Darwin G. Caldwell, Hans Wernher van de Venn
Abstract: The robust interpretation of 3D environments is crucial for human-robot collaboration (HRC) applications, where safety and operational efficiency are paramount. Semantic segmentation plays a key role in this context by enabling a precise and detailed understanding of the environment. Considering the intense data hunger for real-world industrial annotated data essential for effective semantic segmentation, this paper introduces a pioneering approach in the Sim2Real domain adaptation for semantic segmentation of 3D point cloud data, specifically tailored for HRC. Our focus is on developing a network that robustly transitions from simulated environments to real-world applications, thereby enhancing its practical utility and impact on a safe HRC. In this work, we propose a dual-stream network architecture (FUSION) combining Dynamic Graph Convolutional Neural Networks (DGCNN) and Convolutional Neural Networks (CNN) augmented with residual layers as a Sim2Real domain adaptation algorithm for an industrial environment. The proposed model was evaluated on real-world HRC setups and simulation industrial point clouds, it showed increased state-of-the-art performance, achieving a segmentation accuracy of 97.76%, and superior robustness compared to existing methods.
Authors: Weiying Zheng, Ziyue Lin, Pengxin Guo, Yuyin Zhou, Feifei Wang, Liangqiong Qu
Abstract: Vision-Language Models (VLMs) have demonstrated remarkable capabilities in cross-modal understanding and generation by integrating visual and textual information. While instruction tuning and parameter-efficient fine-tuning methods have substantially improved the generalization of VLMs, most existing approaches rely on centralized training, posing challenges for deployment in domains with strict privacy requirements like healthcare. Recent efforts have introduced Federated Learning (FL) into VLM fine-tuning to address these privacy concerns, yet comprehensive benchmarks for evaluating federated fine-tuning strategies, model architectures, and task generalization remain lacking. In this work, we present \textbf{FedVLMBench}, the first systematic benchmark for federated fine-tuning of VLMs. FedVLMBench integrates two mainstream VLM architectures (encoder-based and encoder-free), four fine-tuning strategies, five FL algorithms, six multimodal datasets spanning four cross-domain single-task scenarios and two cross-domain multitask settings, covering four distinct downstream task categories. Through extensive experiments, we uncover key insights into the interplay between VLM architectures, fine-tuning strategies, data heterogeneity, and multi-task federated optimization. Notably, we find that a 2-layer multilayer perceptron (MLP) connector with concurrent connector and LLM tuning emerges as the optimal configuration for encoder-based VLMs in FL. Furthermore, current FL methods exhibit significantly higher sensitivity to data heterogeneity in vision-centric tasks than text-centric ones, across both encoder-free and encoder-based VLM architectures. Our benchmark provides essential tools, datasets, and empirical guidance for the research community, offering a standardized platform to advance privacy-preserving, federated training of multimodal foundation models.
Authors: Harry Walsh, Maksym Ivashechkin, Richard Bowden
Abstract: Machine learning models fundamentally rely on large quantities of high-quality data. Collecting the necessary data for these models can be challenging due to cost, scarcity, and privacy restrictions. Signed languages are visual languages used by the deaf community and are considered low-resource languages. Sign language datasets are often orders of magnitude smaller than their spoken language counterparts. Sign Language Production is the task of generating sign language videos from spoken language sentences, while Sign Language Translation is the reverse translation task. Here, we propose leveraging recent advancements in Sign Language Production to augment existing sign language datasets and enhance the performance of Sign Language Translation models. For this, we utilize three techniques: a skeleton-based approach to production, sign stitching, and two photo-realistic generative models, SignGAN and SignSplat. We evaluate the effectiveness of these techniques in enhancing the performance of Sign Language Translation models by generating variation in the signer's appearance and the motion of the skeletal data. Our results demonstrate that the proposed methods can effectively augment existing datasets and enhance the performance of Sign Language Translation models by up to 19%, paving the way for more robust and accurate Sign Language Translation systems, even in resource-constrained environments.
Authors: Garima Jain, Sanghamitra Pati, Mona Duggal, Amit Sethi, Abhijeet Patil, Gururaj Malekar, Nilesh Kowe, Jitender Kumar, Jatin Kashyap, Divyajeet Rout, Deepali, Hitesh, Nishi Halduniya, Sharat Kumar, Heena Tabassum, Rupinder Singh Dhaliwal, Sucheta Devi Khuraijam, Sushma Khuraijam, Sharmila Laishram, Simmi Kharb, Sunita Singh, K. Swaminadtan, Ranjana Solanki, Deepika Hemranjani, Shashank Nath Singh, Uma Handa, Manveen Kaur, Surinder Singhal, Shivani Kalhan, Rakesh Kumar Gupta, Ravi. S, D. Pavithra, Sunil Kumar Mahto, Arvind Kumar, Deepali Tirkey, Saurav Banerjee, L. Sreelakshmi
Abstract: Oral squamous cell carcinoma OSCC is a major global health burden, particularly in several regions across Asia, Africa, and South America, where it accounts for a significant proportion of cancer cases. Early detection dramatically improves outcomes, with stage I cancers achieving up to 90 percent survival. However, traditional diagnosis based on histopathology has limited accessibility in low-resource settings because it is invasive, resource-intensive, and reliant on expert pathologists. On the other hand, oral cytology of brush biopsy offers a minimally invasive and lower cost alternative, provided that the remaining challenges, inter observer variability and unavailability of expert pathologists can be addressed using artificial intelligence. Development and validation of robust AI solutions requires access to large, labeled, and multi-source datasets to train high capacity models that generalize across domain shifts. We introduce the first large and multicenter oral cytology dataset, comprising annotated slides stained with Papanicolaou(PAP) and May-Grunwald-Giemsa(MGG) protocols, collected from ten tertiary medical centers in India. The dataset is labeled and annotated by expert pathologists for cellular anomaly classification and detection, is designed to advance AI driven diagnostic methods. By filling the gap in publicly available oral cytology datasets, this resource aims to enhance automated detection, reduce diagnostic errors, and improve early OSCC diagnosis in resource-constrained settings, ultimately contributing to reduced mortality and better patient outcomes worldwide.
Authors: Jacob Munkberg, Zian Wang, Ruofan Liang, Tianchang Shen, Jon Hasselgren
Abstract: We leverage finetuned video diffusion models, intrinsic decomposition of videos, and physically-based differentiable rendering to generate high quality materials for 3D models given a text prompt or a single image. We condition a video diffusion model to respect the input geometry and lighting condition. This model produces multiple views of a given 3D model with coherent material properties. Secondly, we use a recent model to extract intrinsics (base color, roughness, metallic) from the generated video. Finally, we use the intrinsics alongside the generated video in a differentiable path tracer to robustly extract PBR materials directly compatible with common content creation tools.
Authors: Alexander Lobashev, Assel Yermekova, Maria Larchenko
Abstract: This paper introduces Factorized MKL-VC, a training-free modification for kNN-VC pipeline. In contrast with original pipeline, our algorithm performs high quality any-to-any cross-lingual voice conversion with only 5 second of reference audio. MKL-VC replaces kNN regression with a factorized optimal transport map in WavLM embedding subspaces, derived from Monge-Kantorovich Linear solution. Factorization addresses non-uniform variance across dimensions, ensuring effective feature transformation. Experiments on LibriSpeech and FLEURS datasets show MKL-VC significantly improves content preservation and robustness with short reference audio, outperforming kNN-VC. MKL-VC achieves performance comparable to FACodec, especially in cross-lingual voice conversion domain.
Authors: Minjong Cheon
Abstract: The advent of Large Weather Models (LWMs) has marked a turning point in data-driven forecasting, with many models now outperforming traditional numerical systems in the medium range. However, achieving stable, long-range autoregressive forecasts beyond a few weeks remains a significant challenge. Prevailing state-of-the-art models that achieve year-long stability, such as SFNO and DLWP-HPX, have relied on transforming input data onto non-standard spatial domains like spherical harmonics or HEALPix meshes. This has led to the prevailing assumption that such representations are necessary to enforce physical consistency and long-term stability. This paper challenges that assumption by investigating whether comparable long-range performance can be achieved on the standard latitude-longitude grid. We introduce AtmosMJ, a deep convolutional network that operates directly on ERA5 data without any spherical remapping. The model's stability is enabled by a novel Gated Residual Fusion (GRF) mechanism, which adaptively moderates feature updates to prevent error accumulation over long recursive simulations. Our results demonstrate that AtmosMJ produces stable and physically plausible forecasts for about 500 days. In quantitative evaluations, it achieves competitive 10-day forecast accuracy against models like Pangu-Weather and GraphCast, all while requiring a remarkably low training budget of 5.7 days on a V100 GPU. Our findings suggest that efficient architectural design, rather than non-standard data representation, can be the key to unlocking stable and computationally efficient long-range weather prediction.
Authors: Zhenran Xu, Yiyu Wang, Xue Yang, Longyue Wang, Weihua Luo, Kaifu Zhang, Baotian Hu, Min Zhang
Abstract: AI-generated content has evolved from monolithic models to modular workflows, particularly on platforms like ComfyUI, enabling customization in creative pipelines. However, crafting effective workflows requires great expertise to orchestrate numerous specialized components, presenting a steep learning curve for users. To address this challenge, we introduce ComfyUI-R1, the first large reasoning model for automated workflow generation. Starting with our curated dataset of 4K workflows, we construct long chain-of-thought (CoT) reasoning data, including node selection, workflow planning, and code-level workflow representation. ComfyUI-R1 is trained through a two-stage framework: (1) CoT fine-tuning for cold start, adapting models to the ComfyUI domain; (2) reinforcement learning for incentivizing reasoning capability, guided by a fine-grained rule-metric hybrid reward, ensuring format validity, structural integrity, and node-level fidelity. Experiments show that our 7B-parameter model achieves a 97\% format validity rate, along with high pass rate, node-level and graph-level F1 scores, significantly surpassing prior state-of-the-art methods that employ leading closed-source models such as GPT-4o and Claude series. Further analysis highlights the critical role of the reasoning process and the advantage of transforming workflows into code. Qualitative comparison reveals our strength in synthesizing intricate workflows with diverse nodes, underscoring the potential of long CoT reasoning in AI art creation.
Authors: Tomas Peterka, Matyas Bohacek
Abstract: Out-of-context and misattributed imagery is the leading form of media manipulation in today's misinformation and disinformation landscape. The existing methods attempting to detect this practice often only consider whether the semantics of the imagery corresponds to the text narrative, missing manipulation so long as the depicted objects or scenes somewhat correspond to the narrative at hand. To tackle this, we introduce News Media Provenance Dataset, a dataset of news articles with provenance-tagged images. We formulate two tasks on this dataset, location of origin relevance (LOR) and date and time of origin relevance (DTOR), and present baseline results on six large language models (LLMs). We identify that, while the zero-shot performance on LOR is promising, the performance on DTOR hinders, leaving room for specialized architectures and future work.
Authors: Irving Fang, Juexiao Zhang, Shengbang Tong, Chen Feng
Abstract: One promise that Vision-Language-Action (VLA) models hold over traditional imitation learning for robotics is to leverage the broad generalization capabilities of large Vision-Language Models (VLMs) to produce versatile, "generalist" robot policies. However, current evaluations of VLAs remain insufficient. Traditional imitation learning benchmarks are unsuitable due to the lack of language instructions. Emerging benchmarks for VLAs that incorporate language often come with limited evaluation tasks and do not intend to investigate how much VLM pretraining truly contributes to the generalization capabilities of the downstream robotic policy. Meanwhile, much research relies on real-world robot setups designed in isolation by different institutions, which creates a barrier for reproducibility and accessibility. To address this gap, we introduce a unified probing suite of 50 simulation-based tasks across 10 subcategories spanning language instruction, vision, and objects. We systematically evaluate several state-of-the-art VLA architectures on this suite to understand their generalization capability. Our results show that while VLM backbones endow VLAs with robust perceptual understanding and high level planning, which we refer to as good intentions, this does not reliably translate into precise motor execution: when faced with out-of-distribution observations, policies often exhibit coherent intentions, but falter in action execution. Moreover, finetuning on action data can erode the original VLM's generalist reasoning abilities. We release our task suite and evaluation code to serve as a standardized benchmark for future VLAs and to drive research on closing the perception-to-action gap. More information, including the source code, can be found at https://ai4ce.github.io/INT-ACT/
Authors: Jared Lawson, Rohan Chitale, Nabil Simaan
Abstract: Safe navigation of steerable and robotic catheters in the cerebral vasculature requires awareness of the catheters shape and pose. Currently, a significant perception burden is placed on interventionalists to mentally reconstruct and predict catheter motions from biplane fluoroscopy images. Efforts to track these catheters are limited to planar segmentation or bulky sensing instrumentation, which are incompatible with microcatheters used in neurointervention. In this work, a catheter is equipped with custom radiopaque markers arranged to enable simultaneous shape and pose estimation under biplane fluoroscopy. A design measure is proposed to guide the arrangement of these markers to minimize sensitivity to marker tracking uncertainty. This approach was deployed for microcatheters smaller than 2mm OD navigating phantom vasculature with shape tracking errors less than 1mm and catheter roll errors below 40 degrees. This work can enable steerable catheters to autonomously navigate under biplane imaging.
Authors: Mahrokh Najaf, Gregory Ongie
Abstract: Implicit neural representations (INRs) have emerged as a powerful tool for solving inverse problems in computer vision and computational imaging. INRs represent images as continuous domain functions realized by a neural network taking spatial coordinates as inputs. However, unlike traditional pixel representations, little is known about the sample complexity of estimating images using INRs in the context of linear inverse problems. Towards this end, we study the sampling requirements for recovery of a continuous domain image from its low-pass Fourier samples by fitting a single hidden-layer INR with ReLU activation and a Fourier features layer using a generalized form of weight decay regularization. Our key insight is to relate minimizers of this non-convex parameter space optimization problem to minimizers of a convex penalty defined over an infinite-dimensional space of measures. We identify a sufficient number of Fourier samples for which an image realized by an INR is exactly recoverable by solving the INR training problem. To validate our theory, we empirically assess the probability of achieving exact recovery of images realized by low-width single hidden-layer INRs, and illustrate the performance of INRs on super-resolution recovery of continuous domain phantom images.
Authors: Yitao Xu, Tong Zhang, Ehsan Pajouheshgar, Sabine S\"usstrunk
Abstract: Conditional diffusion models (CDMs) have shown impressive performance across a range of generative tasks. Their ability to model the full data distribution has opened new avenues for analysis-by-synthesis in downstream discriminative learning. However, this same modeling capacity causes CDMs to entangle the class-defining features with irrelevant context, posing challenges to extracting robust and interpretable representations. To this end, we identify Canonical LAtent Representations (CLAReps), latent codes whose internal CDM features preserve essential categorical information while discarding non-discriminative signals. When decoded, CLAReps produce representative samples for each class, offering an interpretable and compact summary of the core class semantics with minimal irrelevant details. Exploiting CLAReps, we develop a novel diffusion-based feature-distillation paradigm, CaDistill. While the student has full access to the training set, the CDM as teacher transfers core class knowledge only via CLAReps, which amounts to merely 10 % of the training data in size. After training, the student achieves strong adversarial robustness and generalization ability, focusing more on the class signals instead of spurious background cues. Our findings suggest that CDMs can serve not just as image generators but also as compact, interpretable teachers that can drive robust representation learning.
Authors: Mido Assran, Adrien Bardes, David Fan, Quentin Garrido, Russell Howes, Mojtaba, Komeili, Matthew Muckley, Ammar Rizvi, Claire Roberts, Koustuv Sinha, Artem Zholus, Sergio Arnaud, Abha Gejji, Ada Martin, Francois Robert Hogan, Daniel Dugas, Piotr Bojanowski, Vasil Khalidov, Patrick Labatut, Francisco Massa, Marc Szafraniec, Kapil Krishnakumar, Yong Li, Xiaodong Ma, Sarath Chandar, Franziska Meier, Yann LeCun, Michael Rabbat, Nicolas Ballas
Abstract: A major challenge for modern AI is to learn to understand the world and learn to act largely by observation. This paper explores a self-supervised approach that combines internet-scale video data with a small amount of interaction data (robot trajectories), to develop models capable of understanding, predicting, and planning in the physical world. We first pre-train an action-free joint-embedding-predictive architecture, V-JEPA 2, on a video and image dataset comprising over 1 million hours of internet video. V-JEPA 2 achieves strong performance on motion understanding (77.3 top-1 accuracy on Something-Something v2) and state-of-the-art performance on human action anticipation (39.7 recall-at-5 on Epic-Kitchens-100) surpassing previous task-specific models. Additionally, after aligning V-JEPA 2 with a large language model, we demonstrate state-of-the-art performance on multiple video question-answering tasks at the 8 billion parameter scale (e.g., 84.0 on PerceptionTest, 76.9 on TempCompass). Finally, we show how self-supervised learning can be applied to robotic planning tasks by post-training a latent action-conditioned world model, V-JEPA 2-AC, using less than 62 hours of unlabeled robot videos from the Droid dataset. We deploy V-JEPA 2-AC zero-shot on Franka arms in two different labs and enable picking and placing of objects using planning with image goals. Notably, this is achieved without collecting any data from the robots in these environments, and without any task-specific training or reward. This work demonstrates how self-supervised learning from web-scale data and a small amount of robot interaction data can yield a world model capable of planning in the physical world.
Authors: Wenbo Zhang, Tianrun Hu, Yanyuan Qiao, Hanbo Zhang, Yuchu Qin, Yang Li, Jiajun Liu, Tao Kong, Lingqiao Liu, Xiao Ma
Abstract: We present Chain-of-Action (CoA), a novel visuo-motor policy paradigm built upon Trajectory Autoregressive Modeling. Unlike conventional approaches that predict next step action(s) forward, CoA generates an entire trajectory by explicit backward reasoning with task-specific goals through an action-level Chain-of-Thought (CoT) process. This process is unified within a single autoregressive structure: (1) the first token corresponds to a stable keyframe action that encodes the task-specific goals; and (2) subsequent action tokens are generated autoregressively, conditioned on the initial keyframe and previously predicted actions. This backward action reasoning enforces a global-to-local structure, allowing each local action to be tightly constrained by the final goal. To further realize the action reasoning structure, CoA incorporates four complementary designs: continuous action token representation; dynamic stopping for variable-length trajectory generation; reverse temporal ensemble; and multi-token prediction to balance action chunk modeling with global structure. As a result, CoA gives strong spatial generalization capabilities while preserving the flexibility and simplicity of a visuo-motor policy. Empirically, we observe CoA achieves the state-of-the-art performance across 60 RLBench tasks and 8 real-world manipulation tasks.
Authors: Chieh Hubert Lin, Zhaoyang Lv, Songyin Wu, Zhen Xu, Thu Nguyen-Phuoc, Hung-Yu Tseng, Julian Straub, Numair Khan, Lei Xiao, Ming-Hsuan Yang, Yuheng Ren, Richard Newcombe, Zhao Dong, Zhengqin Li
Abstract: We introduce the Deformable Gaussian Splats Large Reconstruction Model (DGS-LRM), the first feed-forward method predicting deformable 3D Gaussian splats from a monocular posed video of any dynamic scene. Feed-forward scene reconstruction has gained significant attention for its ability to rapidly create digital replicas of real-world environments. However, most existing models are limited to static scenes and fail to reconstruct the motion of moving objects. Developing a feed-forward model for dynamic scene reconstruction poses significant challenges, including the scarcity of training data and the need for appropriate 3D representations and training paradigms. To address these challenges, we introduce several key technical contributions: an enhanced large-scale synthetic dataset with ground-truth multi-view videos and dense 3D scene flow supervision; a per-pixel deformable 3D Gaussian representation that is easy to learn, supports high-quality dynamic view synthesis, and enables long-range 3D tracking; and a large transformer network that achieves real-time, generalizable dynamic scene reconstruction. Extensive qualitative and quantitative experiments demonstrate that DGS-LRM achieves dynamic scene reconstruction quality comparable to optimization-based methods, while significantly outperforming the state-of-the-art predictive dynamic reconstruction method on real-world examples. Its predicted physically grounded 3D deformation is accurate and can readily adapt for long-range 3D tracking tasks, achieving performance on par with state-of-the-art monocular video 3D tracking methods.
Authors: Brandon Trabucco, Kyle Doherty, Max Gurinas, Ruslan Salakhutdinov
Abstract: Data augmentation is one of the most prevalent tools in deep learning, underpinning many recent advances, including those from classification, generative models, and representation learning. The standard approach to data augmentation combines simple transformations like rotations and flips to generate new images from existing ones. However, these new images lack diversity along key semantic axes present in the data. Current augmentations cannot alter the high-level semantic attributes, such as animal species present in a scene, to enhance the diversity of data. We address the lack of diversity in data augmentation with image-to-image transformations parameterized by pre-trained text-to-image diffusion models. Our method edits images to change their semantics using an off-the-shelf diffusion model, and generalizes to novel visual concepts from a few labelled examples. We evaluate our approach on few-shot image classification tasks, and on a real-world weed recognition task, and observe an improvement in accuracy in tested domains.
Authors: Aniket Dashpute, Jiazhang Wang, James Taylor, Oliver Cossairt, Ashok Veeraraghavan, Florian Willomitzer
Abstract: Event-based structured light systems have recently been introduced as an exciting alternative to conventional frame-based triangulation systems for the 3D measurements of diffuse surfaces. Important benefits include the fast capture speed and the high dynamic range provided by the event camera - albeit at the cost of lower data quality. So far, both low-accuracy event-based and high-accuracy frame-based 3D imaging systems are tailored to a specific surface type, such as diffuse or specular, and can not be used for a broader class of object surfaces ("mixed reflectance scenes"). In this work, we present a novel event-based structured light system that enables fast 3D imaging of mixed reflectance scenes with high accuracy. On the captured events, we use epipolar constraints that intrinsically enable decomposing the measured reflections into diffuse, two-bounce specular, and other multi-bounce reflections. The diffuse surfaces in the scene are reconstructed using triangulation. Then, the reconstructed diffuse scene parts are leveraged as a "display" to evaluate the specular scene parts via deflectometry. This novel procedure allows us to use the entire scene as a virtual screen, using only a scanning laser and an event camera. The resulting system achieves fast and motion-robust (14Hz) reconstructions of mixed reflectance scenes with < 600 ${\mu}m$ depth error. Moreover, we introduce an "ultrafast" capture mode (250Hz) for the 3D measurement of diffuse scenes.
Authors: Houting Li, Mengxuan Dong, Lok Ming Lui
Abstract: Accurate analysis and classification of facial attributes are essential in various applications, from human-computer interaction to security systems. In this work, a novel approach to enhance facial classification and recognition tasks through the integration of 3D facial models with deep learning methods was proposed. We extract the most useful information for various tasks using the 3D Facial Model, leading to improved classification accuracy. Combining 3D facial insights with ResNet architecture, our approach achieves notable results: 100% individual classification, 95.4% gender classification, and 83.5% expression classification accuracy. This method holds promise for advancing facial analysis and recognition research.
Authors: Kanchana Ranasinghe, Xiang Li, Kumara Kahatapitiya, Michael S. Ryoo
Abstract: Large Language Models (LLMs) have allowed recent LLM-based approaches to achieve excellent performance on long-video understanding benchmarks. We investigate how extensive world knowledge and strong reasoning skills of underlying LLMs influence this strong performance. Surprisingly, we discover that LLM-based approaches can yield surprisingly good accuracy on long-video tasks with limited video information, sometimes even with no video specific information. Building on this, we explore injecting video-specific information into an LLM-based framework. We utilize off-the-shelf vision tools to extract three object-centric information modalities from videos, and then leverage natural language as a medium for fusing this information. Our resulting Multimodal Video Understanding (MVU) framework demonstrates state-of-the-art performance across multiple video understanding benchmarks. Strong performance also on robotics domain tasks establish its strong generality. Code: https://github.com/kahnchana/mvu
Authors: Jingqun Tang, Chunhui Lin, Zhen Zhao, Shu Wei, Binghong Wu, Qi Liu, Yangfan He, Kuan Lu, Hao Feng, Yang Li, Siqi Wang, Lei Liao, Wei Shi, Yuliang Liu, Hao Liu, Yuan Xie, Xiang Bai, Can Huang
Abstract: Text-centric visual question answering (VQA) has made great strides with the development of Multimodal Large Language Models (MLLMs), yet open-source models still fall short of leading models like GPT4V and Gemini, partly due to a lack of extensive, high-quality instruction tuning data. To this end, we introduce a new approach for creating a massive, high-quality instruction-tuning dataset, Square-10M, which is generated using closed-source MLLMs. The data construction process, termed Square, consists of four steps: Self-Questioning, Answering, Reasoning, and Evaluation. Our experiments with Square-10M led to three key findings: 1) Our model, TextSquare, considerably surpasses open-source previous state-of-the-art Text-centric MLLMs and sets a new standard on OCRBench(62.2%). It even outperforms top-tier models like GPT4V and Gemini in 6 of 10 text-centric benchmarks. 2) Additionally, we demonstrate the critical role of VQA reasoning data in offering comprehensive contextual insights for specific questions. This not only improves accuracy but also significantly mitigates hallucinations. Specifically, TextSquare scores an average of 75.1% across four general VQA and hallucination evaluation datasets, outperforming previous state-of-the-art models. 3) Notably, the phenomenon observed in scaling text-centric VQA datasets reveals a vivid pattern: the exponential increase of instruction tuning data volume is directly proportional to the improvement in model performance, thereby validating the necessity of the dataset scale and the high quality of Square-10M.
Authors: Jingqun Tang, Qi Liu, Yongjie Ye, Jinghui Lu, Shu Wei, Chunhui Lin, Wanqing Li, Mohamad Fitri Faiz Bin Mahmood, Hao Feng, Zhen Zhao, Yangfan He, Kuan Lu, Yanjie Wang, Yuliang Liu, Hao Liu, Xiang Bai, Can Huang
Abstract: Text-Centric Visual Question Answering (TEC-VQA) in its proper format not only facilitates human-machine interaction in text-centric visual environments but also serves as a de facto gold proxy to evaluate AI models in the domain of text-centric scene understanding. Nonetheless, most existing TEC-VQA benchmarks have focused on high-resource languages like English and Chinese. Despite pioneering works to expand multilingual QA pairs in non-text-centric VQA datasets through translation engines, the translation-based protocol encounters a substantial "visual-textual misalignment" problem when applied to TEC-VQA. Specifically, it prioritizes the text in question-answer pairs while disregarding the visual text present in images. Moreover, it fails to address complexities related to nuanced meaning, contextual distortion, language bias, and question-type diversity. In this work, we tackle multilingual TEC-VQA by introducing MTVQA, the first benchmark featuring high-quality human expert annotations across 9 diverse languages, consisting of 6,778 question-answer pairs across 2,116 images. Further, by comprehensively evaluating numerous state-of-the-art Multimodal Large Language Models~(MLLMs), including Qwen2-VL, GPT-4o, GPT-4V, Claude3, and Gemini, on the MTVQA benchmark, it is evident that there is still a large room for performance improvement (Qwen2-VL scoring 30.9 versus 79.7 for human performance), underscoring the value of MTVQA. Additionally, we supply multilingual training data within the MTVQA dataset, demonstrating that straightforward fine-tuning with this data can substantially enhance multilingual TEC-VQA performance. We aspire that MTVQA will offer the research community fresh insights and stimulate further exploration in multilingual visual text comprehension. The project homepage is available at https://bytedance.github.io/MTVQA/.
Authors: Han Sun, Yunkang Cao, Hao Dong, Olga Fink
Abstract: Visual anomaly detection (AD) presents significant challenges due to the scarcity of anomalous data samples. While numerous works have been proposed to synthesize anomalous samples, these synthetic anomalies often lack authenticity or require extensive training data, limiting their applicability in real-world scenarios. In this work, we propose Anomaly Anything (AnomalyAny), a novel framework that leverages Stable Diffusion (SD)'s image generation capabilities to generate diverse and realistic unseen anomalies. By conditioning on a single normal sample during test time, AnomalyAny is able to generate unseen anomalies for arbitrary object types with text descriptions. Within AnomalyAny, we propose attention-guided anomaly optimization to direct SD attention on generating hard anomaly concepts. Additionally, we introduce prompt-guided anomaly refinement, incorporating detailed descriptions to further improve the generation quality. Extensive experiments on MVTec AD and VisA datasets demonstrate AnomalyAny's ability in generating high-quality unseen anomalies and its effectiveness in enhancing downstream AD performance.
Authors: Sophie Ostmeier, Brian Axelrod, Maya Varma, Michael E. Moseley, Akshay Chaudhari, Curtis Langlotz
Abstract: Transformer architectures depend on explicit position encodings to capture token positional information. Rotary Position Encoding (RoPE) has emerged as a popular choice in language models due to its efficient encoding of relative position information through key-query rotations. However, RoPE faces significant limitations beyond language processing: it is constrained to one-dimensional sequence data and, even with learnable phases, offers limited representational capacity. We address these challenges with Lie Relative Encodings (LieRE), which generalizes RoPE to high-dimensional rotation matrices by leveraging their Lie group structure. Through extensive evaluation on three image datasets across 2D and 3D classification tasks, LieRE achieves 1.5% improvement over state-of-the-art baselines on 2D tasks and 1% on 3D tasks, while demonstrating superior generalization to higher resolutions. Our implementation is computationally efficient, with results reproducible on 4 A100 GPUs in 30 minutes on CIFAR100. Our code is available at https://github.com/StanfordMIMI/LieRE.
Authors: Shilei Cao, Juepeng Zheng, Yan Liu, Baoquan Zhao, Ziqi Yuan, Weijia Li, Runmin Dong, Haohuan Fu
Abstract: Real-world application models are commonly deployed in dynamic environments, where the target domain distribution undergoes temporal changes. Continual Test-Time Adaptation (CTTA) has recently emerged as a promising technique to gradually adapt a source-trained model to continually changing target domains. Despite recent advancements in addressing CTTA, two critical issues remain: 1) Fixed thresholds for pseudo-labeling in existing methodologies lead to low-quality pseudo-labels, as model confidence varies across categories and domains; 2) Stochastic parameter restoration methods for mitigating catastrophic forgetting fail to preserve critical information effectively, due to their intrinsic randomness. To tackle these challenges for detection models in CTTA scenarios, we present AMROD, featuring three core components. Firstly, the object-level contrastive learning module extracts object-level features for contrastive learning to refine the feature representation in the target domain. Secondly, the adaptive monitoring module dynamically skips unnecessary adaptation and updates the category-specific threshold based on predicted confidence scores to enable efficiency and improve the quality of pseudo-labels. Lastly, the adaptive randomized restoration mechanism selectively reset inactive parameters with higher possibilities, ensuring the retention of essential knowledge. We demonstrate the effectiveness of AMROD on four CTTA object detection tasks, where AMROD outperforms existing methods, especially achieving a 3.2 mAP improvement and a 20\% increase in efficiency on the Cityscapes-to-Cityscapes-C CTTA task. The code of this work is available at https://github.com/ShileiCao/AMROD.
Authors: Yang Song, Lin Wang
Abstract: 3D object detection is an important task that has been widely applied in autonomous driving. To perform this task, a new trend is to fuse multi-modal inputs, i.e., LiDAR and camera. Under such a trend, recent methods fuse these two modalities by unifying them in the same 3D space. However, during direct fusion in a unified space, the drawbacks of both modalities (LiDAR features struggle with detailed semantic information and the camera lacks accurate 3D spatial information) are also preserved, diluting semantic and spatial awareness of the final unified representation. To address the issue, this letter proposes a novel bidirectional complementary LiDAR-camera fusion framework, called BiCo-Fusion that can achieve robust semantic- and spatial-aware 3D object detection. The key insight is to fuse LiDAR and camera features in a bidirectional complementary way to enhance the semantic awareness of the LiDAR and the 3D spatial awareness of the camera. The enhanced features from both modalities are then adaptively fused to build a semantic- and spatial-aware unified representation. Specifically, we introduce Pre-Fusion consisting of a Voxel Enhancement Module (VEM) to enhance the semantic awareness of voxel features from 2D camera features and Image Enhancement Module (IEM) to enhance the 3D spatial awareness of camera features from 3D voxel features. We then introduce Unified Fusion (U-Fusion) to adaptively fuse the enhanced features from the last stage to build a unified representation. Extensive experiments demonstrate the superiority of our BiCo-Fusion against the prior arts. Project page: https://t-ys.github.io/BiCo-Fusion/.
Authors: Mohammad Erfan Sadeghi, Arash Fayyazi, Suhas Somashekar, Armin Abdollahi, Massoud Pedram
Abstract: Vision Transformers (ViTs) represent a groundbreaking shift in machine learning approaches to computer vision. Unlike traditional approaches, ViTs employ the self-attention mechanism, which has been widely used in natural language processing, to analyze image patches. Despite their advantages in modeling visual tasks, deploying ViTs on hardware platforms, notably Field-Programmable Gate Arrays (FPGAs), introduces considerable challenges. These challenges stem primarily from the non-linear calculations and high computational and memory demands of ViTs. This paper introduces CHOSEN, a software-hardware co-design framework to address these challenges and offer an automated framework for ViT deployment on the FPGAs in order to maximize performance. Our framework is built upon three fundamental contributions: multi-kernel design to maximize the bandwidth, mainly targeting benefits of multi DDR memory banks, approximate non-linear functions that exhibit minimal accuracy degradation, and efficient use of available logic blocks on the FPGA, and efficient compiler to maximize the performance and memory-efficiency of the computing kernels by presenting a novel algorithm for design space exploration to find optimal hardware configuration that achieves optimal throughput and latency. Compared to the state-of-the-art ViT accelerators, CHOSEN achieves a 1.5x and 1.42x improvement in the throughput on the DeiT-S and DeiT-B models.
Authors: Yuyan Chen, Songzhou Yan, Zhihong Zhu, Zhixu Li, Yanghua Xiao
Abstract: Humor, deeply rooted in societal meanings and cultural details, poses a unique challenge for machines. While advances have been made in natural language processing, real-world humor often thrives in a multi-modal context, encapsulated distinctively by memes. This paper poses a particular emphasis on the impact of multi-images on meme captioning. After that, we introduce the \textsc{XMeCap} framework, a novel approach that adopts supervised fine-tuning and reinforcement learning based on an innovative reward model, which factors in both global and local similarities between visuals and text. Our results, benchmarked against contemporary models, manifest a marked improvement in caption generation for both single-image and multi-image memes, as well as different meme categories. \textsc{XMeCap} achieves an average evaluation score of 75.85 for single-image memes and 66.32 for multi-image memes, outperforming the best baseline by 6.75\% and 8.56\%, respectively. This research not only establishes a new frontier in meme-related studies but also underscores the potential of machines in understanding and generating humor in a multi-modal setting.
Authors: Yunji Seo, Young Sun Choi, Hyun Seung Son, Youngjung Uh
Abstract: 3D Gaussian Splatting (3DGS) and its subsequent works are restricted to specific hardware setups, either on only low-cost or on only high-end configurations. Approaches aimed at reducing 3DGS memory usage enable rendering on low-cost GPU but compromise rendering quality, which fails to leverage the hardware capabilities in the case of higher-end GPU. Conversely, methods that enhance rendering quality require high-end GPU with large VRAM, making such methods impractical for lower-end devices with limited memory capacity. Consequently, 3DGS-based works generally assume a single hardware setup and lack the flexibility to adapt to varying hardware constraints. To overcome this limitation, we propose Flexible Level of Detail (FLoD) for 3DGS. FLoD constructs a multi-level 3DGS representation through level-specific 3D scale constraints, where each level independently reconstructs the entire scene with varying detail and GPU memory usage. A level-by-level training strategy is introduced to ensure structural consistency across levels. Furthermore, the multi-level structure of FLoD allows selective rendering of image regions at different detail levels, providing additional memory-efficient rendering options. To our knowledge, among prior works which incorporate the concept of Level of Detail (LoD) with 3DGS, FLoD is the first to follow the core principle of LoD by offering adjustable options for a broad range of GPU settings. Experiments demonstrate that FLoD provides various rendering options with trade-offs between quality and memory usage, enabling real-time rendering under diverse memory constraints. Furthermore, we show that FLoD generalizes to different 3DGS frameworks, indicating its potential for integration into future state-of-the-art developments.
Authors: Leonid Erlygin, Alexey Zaytsev
Abstract: Accurate uncertainty estimation is a critical challenge in open-set recognition, where a probe biometric sample may belong to an unknown identity. It can be addressed through sample quality estimation via probabilistic embeddings. However, the low variance of probabilistic embedding only partly implies a low identification error probability: an embedding of a sample could be close to several classes in a gallery, thus yielding high uncertainty despite high sample quality. We propose HolUE - a holistic uncertainty estimation method based on a Bayesian probabilistic model; it is aware of two sources of ambiguity in the open-set recognition system: (1) the gallery uncertainty caused by overlapping classes and (2) the uncertainty of embeddings. Challenging open-set recognition datasets, such as IJB-C for the image domain and VoxBlink for the audio domain, serve as a testbed for our method. We also provide a new open-set recognition protocol for the identification of whales and dolphins. In all cases, HolUE better identifies recognition errors than alternative uncertainty estimation methods, including those based solely on sample quality.
Authors: Yifang Men, Yuan Yao, Miaomiao Cui, Liefeng Bo
Abstract: Character video synthesis aims to produce realistic videos of animatable characters within lifelike scenes. As a fundamental problem in the computer vision and graphics community, 3D works typically require multi-view captures for per-case training, which severely limits their applicability of modeling arbitrary characters in a short time. Recent 2D methods break this limitation via pre-trained diffusion models, but they struggle for pose generality and scene interaction. To this end, we propose MIMO, a novel framework which can not only synthesize character videos with controllable attributes (i.e., character, motion and scene) provided by simple user inputs, but also simultaneously achieve advanced scalability to arbitrary characters, generality to novel 3D motions, and applicability to interactive real-world scenes in a unified framework. The core idea is to encode the 2D video to compact spatial codes, considering the inherent 3D nature of video occurrence. Concretely, we lift the 2D frame pixels into 3D using monocular depth estimators, and decompose the video clip to three spatial components (i.e., main human, underlying scene, and floating occlusion) in hierarchical layers based on the 3D depth. These components are further encoded to canonical identity code, structured motion code and full scene code, which are utilized as control signals of synthesis process. The design of spatial decomposed modeling enables flexible user control, complex motion expression, as well as 3D-aware synthesis for scene interactions. Experimental results demonstrate effectiveness and robustness of the proposed method.
Authors: Tong Liu, Zhixin Lai, Jiawen Wang, Gengyuan Zhang, Shuo Chen, Philip Torr, Vera Demberg, Volker Tresp, Jindong Gu
Abstract: Diffusion models have recently achieved remarkable advancements in terms of image quality and fidelity to textual prompts. Concurrently, the safety of such generative models has become an area of growing concern. This work introduces a novel type of jailbreak, which triggers T2I models to generate the image with visual text, where the image and the text, although considered to be safe in isolation, combine to form unsafe content. To systematically explore this phenomenon, we propose a dataset to evaluate the current diffusion-based text-to-image (T2I) models under such jailbreak. We benchmark nine representative T2I models, including two closed-source commercial models. Experimental results reveal a concerning tendency to produce unsafe content: all tested models suffer from such type of jailbreak, with rates of unsafe generation ranging from around 10\% to 70\% where DALLE 3 demonstrates almost the highest unsafety. In real-world scenarios, various filters such as keyword blocklists, customized prompt filters, and NSFW image filters, are commonly employed to mitigate these risks. We evaluate the effectiveness of such filters against our jailbreak and found that, while these filters may be effective for single modality detection, they fail to work against our jailbreak. We also investigate the underlying reason for such jailbreaks, from the perspective of text rendering capability and training data. Our work provides a foundation for further development towards more secure and reliable T2I models. Project page at https://multimodalpragmatic.github.io/.
Authors: Sara Ghazanfari, Alexandre Araujo, Prashanth Krishnamurthy, Siddharth Garg, Farshad Khorrami
Abstract: Multi-modal Large Language Models (MLLMs) have recently exhibited impressive general-purpose capabilities by leveraging vision foundation models to encode the core concepts of images into representations. These are then combined with instructions and processed by the language model to generate high-quality responses. Despite significant progress in enhancing the language component, challenges persist in optimally fusing visual encodings within the language model for task-specific adaptability. Recent research has focused on improving this fusion through modality adaptation modules but at the cost of significantly increased model complexity and training data needs. In this paper, we propose EMMA (Efficient Multi-Modal Adaptation), a lightweight cross-modality module designed to efficiently fuse visual and textual encodings, generating instruction-aware visual representations for the language model. Our key contributions include: (1) an efficient early fusion mechanism that integrates vision and language representations with minimal added parameters (less than 0.2% increase in model size), (2) an in-depth interpretability analysis that sheds light on the internal mechanisms of the proposed method; (3) comprehensive experiments that demonstrate notable improvements on both specialized and general benchmarks for MLLMs. Empirical results show that EMMA boosts performance across multiple tasks by up to 9.3% while significantly improving robustness against hallucinations. Our code is available at https://github.com/SaraGhazanfari/EMMA
Authors: Felix Koulischer, Johannes Deleu, Gabriel Raya, Thomas Demeester, Luca Ambrogioni
Abstract: Negative Prompting (NP) is widely utilized in diffusion models, particularly in text-to-image applications, to prevent the generation of undesired features. In this paper, we show that conventional NP is limited by the assumption of a constant guidance scale, which may lead to highly suboptimal results, or even complete failure, due to the non-stationarity and state-dependence of the reverse process. Based on this analysis, we derive a principled technique called Dynamic Negative Guidance, which relies on a near-optimal time and state dependent modulation of the guidance without requiring additional training. Unlike NP, negative guidance requires estimating the posterior class probability during the denoising process, which is achieved with limited additional computational overhead by tracking the discrete Markov Chain during the generative process. We evaluate the performance of DNG class-removal on MNIST and CIFAR10, where we show that DNG leads to higher safety, preservation of class balance and image quality when compared with baseline methods. Furthermore, we show that it is possible to use DNG with Stable Diffusion to obtain more accurate and less invasive guidance than NP.
Authors: Hao Ju, Shaofei Huang, Si Liu, Zhedong Zheng
Abstract: Existing approaches to drone visual geo-localization predominantly adopt the image-based setting, where a single drone-view snapshot is matched with images from other platforms. Such task formulation, however, underutilizes the inherent video output of the drone and is sensitive to occlusions and viewpoint disparity. To address these limitations, we formulate a new video-based drone geo-localization task and propose the Video2BEV paradigm. This paradigm transforms the video into a Bird's Eye View (BEV), simplifying the subsequent \textbf{inter-platform} matching process. In particular, we employ Gaussian Splatting to reconstruct a 3D scene and obtain the BEV projection. Different from the existing transform methods, \eg, polar transform, our BEVs preserve more fine-grained details without significant distortion. To facilitate the discriminative \textbf{intra-platform} representation learning, our Video2BEV paradigm also incorporates a diffusion-based module for generating hard negative samples. To validate our approach, we introduce UniV, a new video-based geo-localization dataset that extends the image-based University-1652 dataset. UniV features flight paths at $30^\circ$ and $45^\circ$ elevation angles with increased frame rates of up to 10 frames per second (FPS). Extensive experiments on the UniV dataset show that our Video2BEV paradigm achieves competitive recall rates and outperforms conventional video-based methods. Compared to other competitive methods, our proposed approach exhibits robustness at lower elevations with more occlusions.
Authors: Jingming Liu, Yumeng Li, Boyuan Xiao, Yichang Jian, Ziang Qin, Tianjia Shao, Yao-Xiang Ding, Kun Zhou
Abstract: Under pure textual modality, Large Language Models (LLMs) have demonstrated remarkable success in complex reasoning tasks by decomposing them into simpler sub-problems. However, Multimodal Large Language Models (MLLMs) still struggle with some seemingly straightforward visual tasks, such as counting and solving jigsaw puzzles. We argue that these tasks challenge the ability of visual-to-textual conversion, where MLLMs convert visual information perceived from the input scene, to textual information for further reasoning and generating the answer. If the complexity of the visual input is beyond the perceptual capability of the MLLMs, without decomposing this conversion process, simply scaling inference-time reasoning cannot solve the task because it repeatedly encounters the same perceptual bottleneck. We propose an approach, autonomous imagination, to enable MLLMs to iteratively modify visual inputs (e.g. isolating objects, rearranging puzzle pieces) into intermediate visual states, decomposing visual-to-textual conversion into closed-loop visual modification steps. We show that, without any retraining, MLLMs can now solve tasks initially beyond their perceptual capability, highlighting that closed-loop visual modification can be an effective way of decomposing the visual reasoning task into solvable substeps. Project page: https://future-item.github.io/autoimagine-site/
Authors: Carlos Esteves, Mohammed Suhail, Ameesh Makadia
Abstract: Image tokenizers map images to sequences of discrete tokens, and are a crucial component of autoregressive transformer-based image generation. The tokens are typically associated with spatial locations in the input image, arranged in raster scan order, which is not ideal for autoregressive modeling. In this paper, we propose to tokenize the image spectrum instead, obtained from a discrete wavelet transform (DWT), such that the sequence of tokens represents the image in a coarse-to-fine fashion. Our tokenizer brings several advantages: 1) it leverages that natural images are more compressible at high frequencies, 2) it can take and reconstruct images of different resolutions without retraining, 3) it improves the conditioning for next-token prediction -- instead of conditioning on a partial line-by-line reconstruction of the image, it takes a coarse reconstruction of the full image, 4) it enables partial decoding where the first few generated tokens can reconstruct a coarse version of the image, 5) it enables autoregressive models to be used for image upsampling. We evaluate the tokenizer reconstruction metrics as well as multiscale image generation, text-guided image upsampling and editing.
Authors: Amit Agarwal, Srikant Panda, Angeline Charles, Bhargava Kumar, Hitesh Patel, Priyaranjan Pattnayak, Taki Hasan Rafi, Tejaswini Kumar, Hansa Meghwani, Karan Gupta, Dong-Kyu Chae
Abstract: Multimodal Large Language Models (MLLMs), are recent advancement of Vision-Language Models (VLMs) that have driven major advances in video understanding. However, their vulnerability to adversarial tampering and manipulations remains underexplored. To address this gap, we introduce \textbf{MVTamperBench}, a benchmark that systematically evaluates MLLM robustness against five prevalent tampering techniques: rotation, masking, substitution, repetition, and dropping; based on real-world visual tampering scenarios such as surveillance interference, social media content edits, and misinformation injection. MVTamperBench comprises ~3.4K original videos, expanded into over ~17K tampered clips covering 19 distinct video manipulation tasks. This benchmark challenges models to detect manipulations in spatial and temporal coherence. We evaluate 45 recent MLLMs from 15+ model families. We reveal substantial variability in resilience across tampering types and show that larger parameter counts do not necessarily guarantee robustness. MVTamperBench sets a new benchmark for developing tamper-resilient MLLM in safety-critical applications, including detecting clickbait, preventing harmful content distribution, and enforcing policies on media platforms. We release all code, data, and benchmark to foster open research in trustworthy video understanding. Code: https://amitbcp.github.io/MVTamperBench/ Data: https://huggingface.co/datasets/Srikant86/MVTamperBench
URLs: https://amitbcp.github.io/MVTamperBench/, https://huggingface.co/datasets/Srikant86/MVTamperBench
Authors: Xindi Wu, Mengzhou Xia, Rulin Shao, Zhiwei Deng, Pang Wei Koh, Olga Russakovsky
Abstract: Training vision-language models via instruction tuning often relies on large mixtures of data spanning diverse tasks and domains. However, these mixtures frequently include redundant information, increasing computational costs without proportional performance gains, necessitating more effective data selection strategies. Existing methods typically rely on task-agnostic heuristics to estimate data importance or focus on optimizing single tasks in isolation, limiting their effectiveness in multitask settings. In this work, we introduce ICONS, a gradient-based Influence CONsensus approach for vision-language data Selection. Our method leverages first-order training dynamics to estimate the influence of individual training examples on validation performance and aggregates these estimates across tasks via majority voting over task-specific influences. This cross-task consensus identifies data points that are consistently valuable across tasks, enabling us to prioritize examples that drive overall performance. The voting-based design further mitigates issues such as score calibration and outlier sensitivity, resulting in robust and scalable data selection for diverse multitask mixtures. With only 20% of the data from LLaVA-665K and Cambrian-7M, our selected subsets retain 98.6% and 98.8% of the performance achieved with full datasets, and can even surpass full data training at a 60% selection ratio on LLaVA-665K. Our approach also generalizes to unseen tasks and architectures, demonstrating strong transfer. We release two compact, high-utility subsets, LLaVA-ICONS-133K and Cambrian-ICONS-1.4M, preserving impactful training examples for efficient and scalable vision-language model development.
Authors: Yangfan He, Sida Li, Jianhui Wang, Kun Li, Xinyuan Song, Xinhang Yuan, Keqin Li, Kuan Lu, Menghao Huo, Jingqun Tang, Yi Xin, Jiaqi Chen, Miao Zhang, Xueqian Wang
Abstract: Recent advancements in text-to-image (T2I) generation using diffusion models have enabled cost-effective video-editing applications by leveraging pre-trained models, eliminating the need for resource-intensive training. However, the frame-independence of T2I generation often results in poor temporal consistency. Existing methods address this issue through temporal layer fine-tuning or inference-based temporal propagation, but these approaches suffer from high training costs or limited temporal coherence. To address these challenges, we propose a General and Efficient Adapter (GE-Adapter) that integrates temporal-spatial and semantic consistency with Baliteral DDIM inversion. This framework introduces three key components: (1) Frame-based Temporal Consistency Blocks (FTC Blocks) to capture frame-specific features and enforce smooth inter-frame transitions via temporally-aware loss functions; (2) Channel-dependent Spatial Consistency Blocks (SCD Blocks) employing bilateral filters to enhance spatial coherence by reducing noise and artifacts; and (3) Token-based Semantic Consistency Module (TSC Module) to maintain semantic alignment using shared prompt tokens and frame-specific tokens. Our method significantly improves perceptual quality, text-image alignment, and temporal coherence, as demonstrated on the MSR-VTT dataset. Additionally, it achieves enhanced fidelity and frame-to-frame coherence, offering a practical solution for T2V editing.
Authors: Longtao Jiang, Zhendong Wang, Jianmin Bao, Wengang Zhou, Dongdong Chen, Lei Shi, Dong Chen, Houqiang Li
Abstract: Object removal has so far been dominated by the mask-and-inpaint paradigm, where the masked region is excluded from the input, leaving models relying on unmasked areas to inpaint the missing region. However, this approach lacks contextual information for the masked area, often resulting in unstable performance. In this work, we introduce SmartEraser, built with a new removing paradigm called Masked-Region Guidance. This paradigm retains the masked region in the input, using it as guidance for the removal process. It offers several distinct advantages: (a) it guides the model to accurately identify the object to be removed, preventing its regeneration in the output; (b) since the user mask often extends beyond the object itself, it aids in preserving the surrounding context in the final result. Leveraging this new paradigm, we present Syn4Removal, a large-scale object removal dataset, where instance segmentation data is used to copy and paste objects onto images as removal targets, with the original images serving as ground truths. Experimental results demonstrate that SmartEraser significantly outperforms existing methods, achieving superior performance in object removal, especially in complex scenes with intricate compositions.
Authors: Shuai Lyu, Zijing Tian, Zhonghong Ou, Yifan Zhu, Xiao Zhang, Qiankun Ha, Haoran Luo, Meina Song
Abstract: Cross-modal retrieval maps data under different modality via semantic relevance. Existing approaches implicitly assume that data pairs are well-aligned and ignore the widely existing annotation noise, i.e., noisy correspondence (NC). Consequently, it inevitably causes performance degradation. Despite attempts that employ the co-teaching paradigm with identical architectures to provide distinct data perspectives, the differences between these architectures are primarily stemmed from random initialization. Thus, the model becomes increasingly homogeneous along with the training process. Consequently, the additional information brought by this paradigm is severely limited. In order to resolve this problem, we introduce a Tripartite learning with Semantic Variation Consistency (TSVC) for robust image-text retrieval. We design a tripartite cooperative learning mechanism comprising a Coordinator, a Master, and an Assistant model. The Coordinator distributes data, and the Assistant model supports the Master model's noisy label prediction with diverse data. Moreover, we introduce a soft label estimation method based on mutual information variation, which quantifies the noise in new samples and assigns corresponding soft labels. We also present a new loss function to enhance robustness and optimize training effectiveness. Extensive experiments on three widely used datasets demonstrate that, even at increasing noise ratios, TSVC exhibits significant advantages in retrieval accuracy and maintains stable training performance.
Authors: Long Peng, Xin Di, Zhanfeng Feng, Wenbo Li, Renjing Pei, Yang Wang, Xueyang Fu, Yang Cao, Zheng-Jun Zha
Abstract: Image restoration aims to recover details and enhance contrast in degraded images. With the growing demand for high-quality imaging (\textit{e.g.}, 4K and 8K), achieving a balance between restoration quality and computational efficiency has become increasingly critical. Existing methods, primarily based on CNNs, Transformers, or their hybrid approaches, apply uniform deep representation extraction across the image. However, these methods often struggle to effectively model long-range dependencies and largely overlook the spatial characteristics of image degradation (regions with richer textures tend to suffer more severe damage), making it hard to achieve the best trade-off between restoration quality and efficiency. To address these issues, we propose a novel texture-aware image restoration method, TAMambaIR, which simultaneously perceives image textures and achieves a trade-off between performance and efficiency. Specifically, we introduce a novel Texture-Aware State Space Model, which enhances texture awareness and improves efficiency by modulating the transition matrix of the state-space equation and focusing on regions with complex textures. Additionally, we design a {Multi-Directional Perception Block} to improve multi-directional receptive fields while maintaining low computational overhead. Extensive experiments on benchmarks for image super-resolution, deraining, and low-light image enhancement demonstrate that TAMambaIR achieves state-of-the-art performance with significantly improved efficiency, establishing it as a robust and efficient framework for image restoration.
Authors: Mozes Jacobs, Roberto C. Budzinski, Lyle Muller, Demba Ba, T. Anderson Keller
Abstract: Traveling waves of neural activity are widely observed in the brain, but their precise computational function remains unclear. One prominent hypothesis is that they enable the transfer and integration of spatial information across neural populations. However, few computational models have explored how traveling waves might be harnessed to perform such integrative processing. Drawing inspiration from the famous "Can one hear the shape of a drum?" problem -- which highlights how normal modes of wave dynamics encode geometric information -- we investigate whether similar principles can be leveraged in artificial neural networks. Specifically, we introduce convolutional recurrent neural networks that learn to produce traveling waves in their hidden states in response to visual stimuli, enabling spatial integration. By then treating these wave-like activation sequences as visual representations themselves, we obtain a powerful representational space that outperforms local feed-forward networks on tasks requiring global spatial context. In particular, we observe that traveling waves effectively expand the receptive field of locally connected neurons, supporting long-range encoding and communication of information. We demonstrate that models equipped with this mechanism solve visual semantic segmentation tasks demanding global integration, significantly outperforming local feed-forward models and rivaling non-local U-Net models with fewer parameters. As a first step toward traveling-wave-based communication and visual representation in artificial networks, our findings suggest wave-dynamics may provide efficiency and training stability benefits, while simultaneously offering a new framework for connecting models to biological recordings of neural activity.
Authors: Danae S\'anchez Villegas, Ingo Ziegler, Desmond Elliott
Abstract: Reasoning over sequences of images remains a challenge for multimodal large language models (MLLMs). While recent models incorporate multi-image data during pre-training, they still struggle to recognize sequential structures, often treating images independently. This work introduces ImageChain, a framework that enhances MLLMs with sequential reasoning capabilities over image data by modeling visual sequences as a multi-turn conversation. In ImageChain, images are interleaved with corresponding textual descriptions to form a controlled dialogue that explicitly captures temporal dependencies and narrative progression. Our method optimizes for the task of next-scene description, where the model generates a context-aware description of an upcoming scene based on preceding visual and textual cues. We demonstrate that our approach improves performance on the next-scene description task -- achieving an average improvement from 3.7% to 19% in SimRate, a metric that quantifies semantic similarity to human-annotated ground truths. Moreover, ImageChain achieves robust zero-shot out-of-domain performance in applications ranging from comics to robotics. Extensive experiments validate that instruction-tuning in a multimodal, multi-turn conversation design is key to bridging the gap between static image understanding and temporally-aware reasoning.
Authors: Ziyi Zhang, Zhen Sun, Zongmin Zhang, Jihui Guo, Xinlei He
Abstract: Multimodal Large Language Models (MLLMs) have become powerful and widely adopted in some practical applications. However, recent research has revealed their vulnerability to multimodal jailbreak attacks, whereby the model can be induced to generate harmful content, leading to safety risks. Although most MLLMs have undergone safety alignment, recent research shows that the visual modality is still vulnerable to jailbreak attacks. In our work, we discover that by using flowcharts with partially harmful information, MLLMs can be induced to provide additional harmful details. Based on this, we propose a jailbreak attack method based on auto-generated flowcharts, FC-Attack. Specifically, FC-Attack first fine-tunes a pre-trained LLM to create a step-description generator based on benign datasets. The generator is then used to produce step descriptions corresponding to a harmful query, which are transformed into flowcharts in 3 different shapes (vertical, horizontal, and S-shaped) as visual prompts. These flowcharts are then combined with a benign textual prompt to execute the jailbreak attack on MLLMs. Our evaluations on Advbench show that FC-Attack attains an attack success rate of up to 96% via images and up to 78% via videos across multiple MLLMs. Additionally, we investigate factors affecting the attack performance, including the number of steps and the font styles in the flowcharts. We also find that FC-Attack can improve the jailbreak performance from 4% to 28% in Claude-3.5 by changing the font style. To mitigate the attack, we explore several defenses and find that AdaShield can largely reduce the jailbreak performance but with the cost of utility drop.
Authors: Christopher Wewer, Bart Pogodzinski, Bernt Schiele, Jan Eric Lenssen
Abstract: We introduce Spatial Reasoning Models (SRMs), a framework to perform reasoning over sets of continuous variables via denoising generative models. SRMs infer continuous representations on a set of unobserved variables, given observations on observed variables. Current generative models on spatial domains, such as diffusion and flow matching models, often collapse to hallucination in case of complex distributions. To measure this, we introduce a set of benchmark tasks that test the quality of complex reasoning in generative models and can quantify hallucination. The SRM framework allows to report key findings about importance of sequentialization in generation, the associated order, as well as the sampling strategies during training. It demonstrates, for the first time, that order of generation can successfully be predicted by the denoising network itself. Using these findings, we can increase the accuracy of specific reasoning tasks from <1% to >50%. Our project website provides additional videos, code, and the benchmark datasets: https://geometric-rl.mpi-inf.mpg.de/srm
Authors: Hongyeob Kim, Inyoung Jung, Dayoon Suh, Youjia Zhang, Sangmin Lee, Sungeun Hong
Abstract: Audio-Visual Question Answering (AVQA) requires not only question-based multimodal reasoning but also precise temporal grounding to capture subtle dynamics for accurate prediction. However, existing methods mainly use question information implicitly, limiting focus on question-specific details. Furthermore, most studies rely on uniform frame sampling, which can miss key question-relevant frames. Although recent Top-K frame selection methods aim to address this, their discrete nature still overlooks fine-grained temporal details. This paper proposes QA-TIGER, a novel framework that explicitly incorporates question information and models continuous temporal dynamics. Our key idea is to use Gaussian-based modeling to adaptively focus on both consecutive and non-consecutive frames based on the question, while explicitly injecting question information and applying progressive refinement. We leverage a Mixture of Experts (MoE) to flexibly implement multiple Gaussian models, activating temporal experts specifically tailored to the question. Extensive experiments on multiple AVQA benchmarks show that QA-TIGER consistently achieves state-of-the-art performance. Code is available at https://aim-skku.github.io/QA-TIGER/
Authors: Parsa Rahimi, Damien Teney, Sebastien Marcel
Abstract: The increasing reliance on large-scale datasets in machine learning poses significant privacy and ethical challenges, particularly in sensitive domains such as face recognition (FR). Synthetic data generation offers a promising alternative; however, most existing methods depend heavily on external datasets or pre-trained models, increasing complexity and resource demands. In this paper, we introduce AugGen, a self-contained synthetic augmentation technique. AugGen strategically samples from a class-conditional generative model trained exclusively on the target FR dataset, eliminating the need for external resources. Evaluated across 8 FR benchmarks, including IJB-C and IJB-B, our method achieves 1-12% performance improvements, outperforming models trained solely on real data and surpassing state-of-the-art synthetic data generation approaches, while using less real data. Notably, these gains often exceed those from architectural modifications, underscoring the value of synthetic augmentation in data-limited scenarios. Our findings demonstrate that carefully integrated synthetic data can both mitigate privacy constraints and substantially enhance discriminative performance in face recognition. Paper website: https://parsa-ra.github.io/auggen/.
Authors: Mo Zhou, Jianwei Wang, Xuanmeng Zhang, Dylan Campbell, Kai Wang, Long Yuan, Wenjie Zhang, Xuemin Lin
Abstract: This paper studies optical flow estimation, a critical task in motion analysis with applications in autonomous navigation, action recognition, and film production. Traditional optical flow methods require consecutive frames, which are often unavailable due to limitations in data acquisition or real-world scene disruptions. Thus, single-frame optical flow estimation is emerging in the literature. However, existing single-frame approaches suffer from two major limitations: (1) they rely on labeled training data, making them task-specific, and (2) they produce deterministic predictions, failing to capture motion uncertainty. To overcome these challenges, we propose ProbDiffFlow, a training-free framework that estimates optical flow distributions from a single image. Instead of directly predicting motion, ProbDiffFlow follows an estimation-by-synthesis paradigm: it first generates diverse plausible future frames using a diffusion-based model, then estimates motion from these synthesized samples using a pre-trained optical flow model, and finally aggregates the results into a probabilistic flow distribution. This design eliminates the need for task-specific training while capturing multiple plausible motions. Experiments on both synthetic and real-world datasets demonstrate that ProbDiffFlow achieves superior accuracy, diversity, and efficiency, outperforming existing single-image and two-frame baselines.
Authors: Siyuan Yang, Shilin Lu, Shizheng Wang, Meng Hwa Er, Zengwei Zheng, Alex C. Kot
Abstract: This paper explores the promising interplay between spiking neural networks (SNNs) and event-based cameras for privacy-preserving human action recognition (HAR). The unique feature of event cameras in capturing only the outlines of motion, combined with SNNs' proficiency in processing spatiotemporal data through spikes, establishes a highly synergistic compatibility for event-based HAR. Previous studies, however, have been limited by SNNs' ability to process long-term temporal information, essential for precise HAR. In this paper, we introduce two novel frameworks to address this: temporal segment-based SNN (\textit{TS-SNN}) and 3D convolutional SNN (\textit{3D-SNN}). The \textit{TS-SNN} extracts long-term temporal information by dividing actions into shorter segments, while the \textit{3D-SNN} replaces 2D spatial elements with 3D components to facilitate the transmission of temporal information. To promote further research in event-based HAR, we create a dataset, \textit{FallingDetection-CeleX}, collected using the high-resolution CeleX-V event camera $(1280 \times 800)$, comprising 7 distinct actions. Extensive experimental results show that our proposed frameworks surpass state-of-the-art SNN methods on our newly collected dataset and three other neuromorphic datasets, showcasing their effectiveness in handling long-range temporal information for event-based HAR.
Authors: Johannes Jakubik, Felix Yang, Benedikt Blumenstiel, Erik Scheurer, Rocco Sedona, Stefano Maurogiovanni, Jente Bosmans, Nikolaos Dionelis, Valerio Marsocci, Niklas Kopp, Rahul Ramachandran, Paolo Fraccaro, Thomas Brunschwiler, Gabriele Cavallaro, Juan Bernabe-Moreno, Nicolas Long\'ep\'e
Abstract: We present TerraMind, the first any-to-any generative, multimodal foundation model for Earth observation (EO). Unlike other multimodal models, TerraMind is pretrained on dual-scale representations combining both token-level and pixel-level data across modalities. On a token level, TerraMind encodes high-level contextual information to learn cross-modal relationships, while on a pixel level, TerraMind leverages fine-grained representations to capture critical spatial nuances. We pretrained TerraMind on nine geospatial modalities of a global, large-scale dataset. In this paper, we demonstrate that (i) TerraMind's dual-scale early fusion approach unlocks a range of zero-shot and few-shot applications for Earth observation, (ii) TerraMind introduces "Thinking-in-Modalities" (TiM) -- the capability of generating additional artificial data during finetuning and inference to improve the model output -- and (iii) TerraMind achieves beyond state-of-the-art performance in community-standard benchmarks for EO like PANGAEA. The pretraining dataset, the model weights, and our code are open-sourced under a permissive license.
Authors: Rezowan Shuvo, M S Mekala, Eyad Elyan
Abstract: Understanding actions within surgical workflows is critical for evaluating post-operative outcomes and enhancing surgical training and efficiency. Capturing and analyzing long sequences of actions in surgical settings is challenging due to the inherent variability in individual surgeon approaches, which are shaped by their expertise and preferences. This variability complicates the identification and segmentation of distinct actions with ambiguous boundary start and end points. The traditional models, such as MS-TCN, which rely on large receptive fields, that causes over-segmentation, or under-segmentation, where distinct actions are incorrectly aligned. To address these challenges, we propose the Multi-Stage Boundary-Aware Transformer Network (MSBATN) with hierarchical sliding window attention to improve action segmentation. Our approach effectively manages the complexity of varying action durations and subtle transitions by accurately identifying start and end action boundaries in untrimmed surgical videos. MSBATN introduces a novel unified loss function that optimises action classification and boundary detection as interconnected tasks. Unlike conventional binary boundary detection methods, our innovative boundary weighing mechanism leverages contextual information to precisely identify action boundaries. Extensive experiments on three challenging surgical datasets demonstrate that MSBATN achieves state-of-the-art performance, with superior F1 scores at 25% and 50%. thresholds and competitive results across other metrics.
Authors: Gaozheng Pei, Ke Ma, Yingfei Sun, Qianqian Xu, Qingming Huang
Abstract: The diffusion-based adversarial purification methods attempt to drown adversarial perturbations into a part of isotropic noise through the forward process, and then recover the clean images through the reverse process. Due to the lack of distribution information about adversarial perturbations in the pixel domain, it is often unavoidable to damage normal semantics. We turn to the frequency domain perspective, decomposing the image into amplitude spectrum and phase spectrum. We find that for both spectra, the damage caused by adversarial perturbations tends to increase monotonically with frequency. This means that we can extract the content and structural information of the original clean sample from the frequency components that are less damaged. Meanwhile, theoretical analysis indicates that existing purification methods indiscriminately damage all frequency components, leading to excessive damage to the image. Therefore, we propose a purification method that can eliminate adversarial perturbations while maximizing the preservation of the content and structure of the original image. Specifically, at each time step during the reverse process, for the amplitude spectrum, we replace the low-frequency components of the estimated image's amplitude spectrum with the corresponding parts of the adversarial image. For the phase spectrum, we project the phase of the estimated image into a designated range of the adversarial image's phase spectrum, focusing on the low frequencies. Empirical evidence from extensive experiments demonstrates that our method significantly outperforms most current defense methods.
Authors: Shang Zhang, Huanbin Zhang, Dali Feng, Yujie Cui, Ruoyan Xiong, Cen He
Abstract: Thermal infrared (TIR) object tracking often suffers from challenges such as target occlusion, motion blur, and background clutter, which significantly degrade the performance of trackers. To address these issues, this paper pro-poses a novel Siamese Motion Mamba Tracker (SMMT), which integrates a bidirectional state-space model and a self-attention mechanism. Specifically, we introduce the Motion Mamba module into the Siamese architecture to ex-tract motion features and recover overlooked edge details using bidirectional modeling and self-attention. We propose a Siamese parameter-sharing strate-gy that allows certain convolutional layers to share weights. This approach reduces computational redundancy while preserving strong feature represen-tation. In addition, we design a motion edge-aware regression loss to improve tracking accuracy, especially for motion-blurred targets. Extensive experi-ments are conducted on four TIR tracking benchmarks, including LSOTB-TIR, PTB-TIR, VOT-TIR2015, and VOT-TIR 2017. The results show that SMMT achieves superior performance in TIR target tracking.
Authors: Woojin Cho, Steve Andreas Immanuel, Junhyuk Heo, Darongsae Kwon
Abstract: Multispectral satellite images play a vital role in agriculture, fisheries, and environmental monitoring. However, their high dimensionality, large data volumes, and diverse spatial resolutions across multiple channels pose significant challenges for data compression and analysis. This paper presents ImpliSat, a unified framework specifically designed to address these challenges through efficient compression and reconstruction of multispectral satellite data. ImpliSat leverages Implicit Neural Representations (INR) to model satellite images as continuous functions over coordinate space, capturing fine spatial details across varying spatial resolutions. Furthermore, we introduce a Fourier modulation algorithm that dynamically adjusts to the spectral and spatial characteristics of each band, ensuring optimal compression while preserving critical image details.
Authors: Martin JJ. Bucher, Iro Armeni
Abstract: Scene synthesis and editing has emerged as a promising direction in computer graphics. Current trained approaches for 3D indoor scenes either oversimplify object semantics through one-hot class encodings (e.g., 'chair' or 'table'), require masked diffusion for editing, ignore room boundaries, or rely on floor plan renderings that fail to capture complex layouts. In contrast, LLM-based methods enable richer semantics via natural language (e.g., 'modern studio with light wood furniture') but do not support editing, remain limited to rectangular layouts or rely on weak spatial reasoning from implicit world models. We introduce ReSpace, a generative framework for text-driven 3D indoor scene synthesis and editing using autoregressive language models. Our approach features a compact structured scene representation with explicit room boundaries that frames scene editing as a next-token prediction task. We leverage a dual-stage training approach combining supervised fine-tuning and preference alignment, enabling a specially trained language model for object addition that accounts for user instructions, spatial geometry, object semantics, and scene-level composition. For scene editing, we employ a zero-shot LLM to handle object removal and prompts for addition. We further introduce a novel voxelization-based evaluation that captures fine-grained geometry beyond 3D bounding boxes. Experimental results surpass state-of-the-art on object addition while maintaining competitive results on full scene synthesis.
Authors: Halil Ismail Helvaci, Justin Philip Huber, Jihye Bae, Sen-ching Samson Cheung
Abstract: Stroke rehabilitation often demands precise tracking of patient movements to monitor progress, with complexities of rehabilitation exercises presenting two critical challenges: fine-grained and sub-second (under one-second) action detection. In this work, we propose the High Resolution Temporal Transformer (HRTR), to time-localize and classify high-resolution (fine-grained), sub-second actions in a single-stage transformer, eliminating the need for multi-stage methods and post-processing. Without any refinements, HRTR outperforms state-of-the-art systems on both stroke related and general datasets, achieving Edit Score (ES) of 70.1 on StrokeRehab Video, 69.4 on StrokeRehab IMU, and 88.4 on 50Salads.
Authors: Qiaohui Chu, Haoyu Zhang, Yisen Feng, Meng Liu, Weili Guan, Yaowei Wang, Liqiang Nie
Abstract: In this report, we present a novel three-stage framework developed for the Ego4D Long-Term Action Anticipation (LTA) task. Inspired by recent advances in foundation models, our method consists of three stages: feature extraction, action recognition, and long-term action anticipation. First, visual features are extracted using a high-performance visual encoder. The features are then fed into a Transformer to predict verbs and nouns, with a verb-noun co-occurrence matrix incorporated to enhance recognition accuracy. Finally, the predicted verb-noun pairs are formatted as textual prompts and input into a fine-tuned large language model (LLM) to anticipate future action sequences. Our framework achieves first place in this challenge at CVPR 2025, establishing a new state-of-the-art in long-term action prediction. Our code will be released at https://github.com/CorrineQiu/Ego4D-LTA-Challenge-2025.
URLs: https://github.com/CorrineQiu/Ego4D-LTA-Challenge-2025.
Authors: Di Chang, Mingdeng Cao, Yichun Shi, Bo Liu, Shengqu Cai, Shijie Zhou, Weilin Huang, Gordon Wetzstein, Mohammad Soleymani, Peng Wang
Abstract: Editing images with instructions to reflect non-rigid motions, camera viewpoint shifts, object deformations, human articulations, and complex interactions, poses a challenging yet underexplored problem in computer vision. Existing approaches and datasets predominantly focus on static scenes or rigid transformations, limiting their capacity to handle expressive edits involving dynamic motion. To address this gap, we introduce ByteMorph, a comprehensive framework for instruction-based image editing with an emphasis on non-rigid motions. ByteMorph comprises a large-scale dataset, ByteMorph-6M, and a strong baseline model built upon the Diffusion Transformer (DiT), named ByteMorpher. ByteMorph-6M includes over 6 million high-resolution image editing pairs for training, along with a carefully curated evaluation benchmark ByteMorph-Bench. Both capture a wide variety of non-rigid motion types across diverse environments, human figures, and object categories. The dataset is constructed using motion-guided data generation, layered compositing techniques, and automated captioning to ensure diversity, realism, and semantic coherence. We further conduct a comprehensive evaluation of recent instruction-based image editing methods from both academic and commercial domains.
Authors: Erhang Zhang, Junyi Ma, Yin-Dong Zheng, Yixuan Zhou, Hesheng Wang
Abstract: Locating human-object interaction (HOI) actions within video serves as the foundation for multiple downstream tasks, such as human behavior analysis and human-robot skill transfer. Current temporal action localization methods typically rely on annotated action and object categories of interactions for optimization, which leads to domain bias and low deployment efficiency. Although some recent works have achieved zero-shot temporal action localization (ZS-TAL) with large vision-language models (VLMs), their coarse-grained estimations and open-loop pipelines hinder further performance improvements for temporal interaction localization (TIL). To address these issues, we propose a novel zero-shot TIL approach dubbed EgoLoc to locate the timings of grasp actions for human-object interaction in egocentric videos. EgoLoc introduces a self-adaptive sampling strategy to generate reasonable visual prompts for VLM reasoning. By absorbing both 2D and 3D observations, it directly samples high-quality initial guesses around the possible contact/separation timestamps of HOI according to 3D hand velocities, leading to high inference accuracy and efficiency. In addition, EgoLoc generates closed-loop feedback from visual and dynamic cues to further refine the localization results. Comprehensive experiments on the publicly available dataset and our newly proposed benchmark demonstrate that EgoLoc achieves better temporal interaction localization for egocentric videos compared to state-of-the-art baselines. We will release our code and relevant data as open-source at https://github.com/IRMVLab/EgoLoc.
Authors: Qiuyu Tang, Bonor Ayambem, Mooi Choo Chuah, Aparna Bharati
Abstract: The remarkable image generation capabilities of state-of-the-art diffusion models, such as Stable Diffusion, can also be misused to spread misinformation and plagiarize copyrighted materials. To mitigate the potential risks associated with image editing, current image protection methods rely on adding imperceptible perturbations to images to obstruct diffusion-based editing. A fully successful protection for an image implies that the output of editing attempts is an undesirable, noisy image which is completely unrelated to the reference image. In our experiments with various perturbation-based image protection methods across multiple domains (natural scene images and artworks) and editing tasks (image-to-image generation and style editing), we discover that such protection does not achieve this goal completely. In most scenarios, diffusion-based editing of protected images generates a desirable output image which adheres precisely to the guidance prompt. Our findings suggest that adding noise to images may paradoxically increase their association with given text prompts during the generation process, leading to unintended consequences such as better resultant edits. Hence, we argue that perturbation-based methods may not provide a sufficient solution for robust image protection against diffusion-based editing.
Authors: Youngwan Lee, Kangsan Kim, Kwanyong Park, Ilcahe Jung, Soojin Jang, Seanie Lee, Yong-Ju Lee, Sung Ju Hwang
Abstract: Despite emerging efforts to enhance the safety of Vision-Language Models (VLMs), current approaches face two main shortcomings. 1) Existing safety-tuning datasets and benchmarks only partially consider how image-text interactions can yield harmful content, often overlooking contextually unsafe outcomes from seemingly benign pairs. This narrow coverage leaves VLMs vulnerable to jailbreak attacks in unseen configurations. 2) Prior methods rely primarily on data-centric tuning, with limited architectural innovations to intrinsically strengthen safety. We address these gaps by introducing a holistic safety dataset and benchmark, HoliSafe, that spans all five safe/unsafe image-text combinations, providing a more robust basis for both training and evaluation. We further propose SafeLLaVA, a novel VLM augmented with a learnable safety meta token and a dedicated safety head. The meta token encodes harmful visual cues during training, intrinsically guiding the language model toward safer responses, while the safety head offers interpretable harmfulness classification aligned with refusal rationales. Experiments show that SafeLLaVA, trained on HoliSafe, achieves state-of-the-art safety performance across multiple VLM benchmarks. Additionally, the HoliSafe benchmark itself reveals critical vulnerabilities in existing models. We hope that HoliSafe and SafeLLaVA will spur further research into robust and interpretable VLM safety, expanding future avenues for multimodal alignment.
Authors: Wenfeng Lin, Renjie Chen, Boyuan Liu, Shiyue Yan, Ruoyu Feng, Jiangchuan Wei, Yichen Zhang, Yimeng Zhou, Chao Feng, Jiao Ran, Qi Wu, Zuotao Liu, Mingyu Guo
Abstract: Recent advances in video generation demand increasingly efficient training recipes to mitigate escalating computational costs. In this report, we present ContentV, an 8B-parameter text-to-video model that achieves state-of-the-art performance (85.14 on VBench) after training on 256 x 64GB Neural Processing Units (NPUs) for merely four weeks. ContentV generates diverse, high-quality videos across multiple resolutions and durations from text prompts, enabled by three key innovations: (1) A minimalist architecture that maximizes reuse of pre-trained image generation models for video generation; (2) A systematic multi-stage training strategy leveraging flow matching for enhanced efficiency; and (3) A cost-effective reinforcement learning with human feedback framework that improves generation quality without requiring additional human annotations. All the code and models are available at: https://contentv.github.io.
Authors: Taiga Shinozaki, Tomoki Doi, Amane Watahiki, Satoshi Nishida, Hitomi Yanaka
Abstract: Humans are susceptible to optical illusions, which serve as valuable tools for investigating sensory and cognitive processes. Inspired by human vision studies, research has begun exploring whether machines, such as large vision language models (LVLMs), exhibit similar susceptibilities to visual illusions. However, studies often have used non-abstract images and have not distinguished actual and apparent features, leading to ambiguous assessments of machine cognition. To address these limitations, we introduce a visual question answering (VQA) dataset, categorized into genuine and fake illusions, along with corresponding control images. Genuine illusions present discrepancies between actual and apparent features, whereas fake illusions have the same actual and apparent features even though they look illusory due to the similar geometric configuration. We evaluate the performance of LVLMs for genuine and fake illusion VQA tasks and investigate whether the models discern actual and apparent features. Our findings indicate that although LVLMs may appear to recognize illusions by correctly answering questions about both feature types, they predict the same answers for both Genuine Illusion and Fake Illusion VQA questions. This suggests that their responses might be based on prior knowledge of illusions rather than genuine visual understanding. The dataset is available at https://github.com/ynklab/FILM
Authors: Zonglin Wu, Yule Xue, Xin Wei, Yiren Song
Abstract: As automated attack techniques rapidly advance, CAPTCHAs remain a critical defense mechanism against malicious bots. However, existing CAPTCHA schemes encompass a diverse range of modalities -- from static distorted text and obfuscated images to interactive clicks, sliding puzzles, and logic-based questions -- yet the community still lacks a unified, large-scale, multimodal benchmark to rigorously evaluate their security robustness. To address this gap, we introduce MCA-Bench, a comprehensive and reproducible benchmarking suite that integrates heterogeneous CAPTCHA types into a single evaluation protocol. Leveraging a shared vision-language model backbone, we fine-tune specialized cracking agents for each CAPTCHA category, enabling consistent, cross-modal assessments. Extensive experiments reveal that MCA-Bench effectively maps the vulnerability spectrum of modern CAPTCHA designs under varied attack settings, and crucially offers the first quantitative analysis of how challenge complexity, interaction depth, and model solvability interrelate. Based on these findings, we propose three actionable design principles and identify key open challenges, laying the groundwork for systematic CAPTCHA hardening, fair benchmarking, and broader community collaboration. Datasets and code are available online.
Authors: Ruoxuan Zhang, Jidong Gao, Bin Wen, Hongxia Xie, Chenming Zhang, Hong-Han Shuai, Wen-Huang Cheng
Abstract: Creating recipe images is a key challenge in food computing, with applications in culinary education and multimodal recipe assistants. However, existing datasets lack fine-grained alignment between recipe goals, step-wise instructions, and visual content. We present RecipeGen, the first large-scale, real-world benchmark for recipe-based Text-to-Image (T2I), Image-to-Video (I2V), and Text-to-Video (T2V) generation. RecipeGen contains 26,453 recipes, 196,724 images, and 4,491 videos, covering diverse ingredients, cooking procedures, styles, and dish types. We further propose domain-specific evaluation metrics to assess ingredient fidelity and interaction modeling, benchmark representative T2I, I2V, and T2V models, and provide insights for future recipe generation models. Project page is available now.
Authors: Xiangyu Guo, Zhanqian Wu, Kaixin Xiong, Ziyang Xu, Lijun Zhou, Gangwei Xu, Shaoqing Xu, Haiyang Sun, Bing Wang, Guang Chen, Hangjun Ye, Wenyu Liu, Xinggang Wang
Abstract: We present Genesis, a unified framework for joint generation of multi-view driving videos and LiDAR sequences with spatio-temporal and cross-modal consistency. Genesis employs a two-stage architecture that integrates a DiT-based video diffusion model with 3D-VAE encoding, and a BEV-aware LiDAR generator with NeRF-based rendering and adaptive sampling. Both modalities are directly coupled through a shared latent space, enabling coherent evolution across visual and geometric domains. To guide the generation with structured semantics, we introduce DataCrafter, a captioning module built on vision-language models that provides scene-level and instance-level supervision. Extensive experiments on the nuScenes benchmark demonstrate that Genesis achieves state-of-the-art performance across video and LiDAR metrics (FVD 16.95, FID 4.24, Chamfer 0.611), and benefits downstream tasks including segmentation and 3D detection, validating the semantic fidelity and practical utility of the generated data.
Authors: Xuemei Chen, Huamin Wang, Hangchi Shen, Shukai Duan, Shiping Wen, Tingwen Huang
Abstract: Low energy consumption for 3D object detection is an important research area because of the increasing energy consumption with their wide application in fields such as autonomous driving. The spiking neural networks (SNNs) with low-power consumption characteristics can provide a novel solution for this research. Therefore, we apply SNNs to monocular 3D object detection and propose the SpikeSMOKE architecture in this paper, which is a new attempt for low-power monocular 3D object detection. As we all know, discrete signals of SNNs will generate information loss and limit their feature expression ability compared with the artificial neural networks (ANNs).In order to address this issue, inspired by the filtering mechanism of biological neuronal synapses, we propose a cross-scale gated coding mechanism(CSGC), which can enhance feature representation by combining cross-scale fusion of attentional methods and gated filtering mechanisms.In addition, to reduce the computation and increase the speed of training, we present a novel light-weight residual block that can maintain spiking computing paradigm and the highest possible detection performance. Compared to the baseline SpikeSMOKE under the 3D Object Detection, the proposed SpikeSMOKE with CSGC can achieve 11.78 (+2.82, Easy), 10.69 (+3.2, Moderate), and 10.48 (+3.17, Hard) on the KITTI autonomous driving dataset by AP|R11 at 0.7 IoU threshold, respectively. It is important to note that the results of SpikeSMOKE can significantly reduce energy consumption compared to the results on SMOKE. For example,the energy consumption can be reduced by 72.2% on the hard category, while the detection performance is reduced by only 4%. SpikeSMOKE-L (lightweight) can further reduce the amount of parameters by 3 times and computation by 10 times compared to SMOKE.
Authors: Yizhen Li, Dell Zhang, Xuelong Li, Yiqing Shen
Abstract: Reasoning Segmentation (RS) is a multimodal vision-text task that requires segmenting objects based on implicit text queries, demanding both precise visual perception and vision-text reasoning capabilities. Current RS approaches rely on fine-tuning vision-language models (VLMs) for both perception and reasoning, but their tokenization of images fundamentally disrupts continuous spatial relationships between objects. We introduce DTwinSeger, a novel RS approach that leverages Digital Twin (DT) representation as an intermediate layer to decouple perception from reasoning. Innovatively, DTwinSeger reformulates RS as a two-stage process, where the first transforms the image into a structured DT representation that preserves spatial relationships and semantic properties and then employs a Large Language Model (LLM) to perform explicit reasoning over this representation to identify target objects. We propose a supervised fine-tuning method specifically for LLM with DT representation, together with a corresponding fine-tuning dataset Seg-DT, to enhance the LLM's reasoning capabilities with DT representations. Experiments show that our method can achieve state-of-the-art performance on two image RS benchmarks and three image referring segmentation benchmarks. It yields that DT representation functions as an effective bridge between vision and text, enabling complex multimodal reasoning tasks to be accomplished solely with an LLM.
Authors: Zhengyao Lv, Tianlin Pan, Chenyang Si, Zhaoxi Chen, Wangmeng Zuo, Ziwei Liu, Kwan-Yee K. Wong
Abstract: Multimodal Diffusion Transformers (MM-DiTs) have achieved remarkable progress in text-driven visual generation. However, even state-of-the-art MM-DiT models like FLUX struggle with achieving precise alignment between text prompts and generated content. We identify two key issues in the attention mechanism of MM-DiT, namely 1) the suppression of cross-modal attention due to token imbalance between visual and textual modalities and 2) the lack of timestep-aware attention weighting, which hinder the alignment. To address these issues, we propose \textbf{Temperature-Adjusted Cross-modal Attention (TACA)}, a parameter-efficient method that dynamically rebalances multimodal interactions through temperature scaling and timestep-dependent adjustment. When combined with LoRA fine-tuning, TACA significantly enhances text-image alignment on the T2I-CompBench benchmark with minimal computational overhead. We tested TACA on state-of-the-art models like FLUX and SD3.5, demonstrating its ability to improve image-text alignment in terms of object appearance, attribute binding, and spatial relationships. Our findings highlight the importance of balancing cross-modal attention in improving semantic fidelity in text-to-image diffusion models. Our codes are publicly available at \href{https://github.com/Vchitect/TACA}
Authors: Nick Jiang, Amil Dravid, Alexei Efros, Yossi Gandelsman
Abstract: We investigate the mechanism underlying a previously identified phenomenon in Vision Transformers -- the emergence of high-norm tokens that lead to noisy attention maps. We observe that in multiple models (e.g., CLIP, DINOv2), a sparse set of neurons is responsible for concentrating high-norm activations on outlier tokens, leading to irregular attention patterns and degrading downstream visual processing. While the existing solution for removing these outliers involves retraining models from scratch with additional learned register tokens, we use our findings to create a training-free approach to mitigate these artifacts. By shifting the high-norm activations from our discovered register neurons into an additional untrained token, we can mimic the effect of register tokens on a model already trained without registers. We demonstrate that our method produces cleaner attention and feature maps, enhances performance over base models across multiple downstream visual tasks, and achieves results comparable to models explicitly trained with register tokens. We then extend test-time registers to off-the-shelf vision-language models to improve their interpretability. Our results suggest that test-time registers effectively take on the role of register tokens at test-time, offering a training-free solution for any pre-trained model released without them.
Authors: Zheng Han, Jun Zhou, Jialun Pei, Jing Qin, Yingfang Fan, Qi Dou
Abstract: In augmented reality (AR)-guided surgical navigation, preoperative organ models are superimposed onto the patient's intraoperative anatomy to visualize critical structures such as vessels and tumors. Accurate deformation modeling is essential to maintain the reliability of AR overlays by ensuring alignment between preoperative models and the dynamically changing anatomy. Although the finite element method (FEM) offers physically plausible modeling, its high computational cost limits intraoperative applicability. Moreover, existing algorithms often fail to handle large anatomical changes, such as those induced by pneumoperitoneum or ligament dissection, leading to inaccurate anatomical correspondences and compromised AR guidance. To address these challenges, we propose a data-driven biomechanics algorithm that preserves FEM-level accuracy while improving computational efficiency. In addition, we introduce a novel human-in-the-loop mechanism into the deformation modeling process. This enables surgeons to interactively provide prompts to correct anatomical misalignments, thereby incorporating clinical expertise and allowing the model to adapt dynamically to complex surgical scenarios. Experiments on a publicly available dataset demonstrate that our algorithm achieves a mean target registration error of 3.42 mm. Incorporating surgeon prompts through the interactive framework further reduces the error to 2.78 mm, surpassing state-of-the-art methods in volumetric accuracy. These results highlight the ability of our framework to deliver efficient and accurate deformation modeling while enhancing surgeon-algorithm collaboration, paving the way for safer and more reliable computer-assisted surgeries.
Authors: Oishee Bintey Hoque, Abhijin Adiga, Aniruddha Adiga, Siddharth Chaudhary, Madhav V. Marathe, S. S. Ravi, Kirti Rajagopalan, Amanda Wilson, Samarth Swarup
Abstract: Accurate canal network mapping is essential for water management, including irrigation planning and infrastructure maintenance. State-of-the-art semantic segmentation models for infrastructure mapping, such as roads, rely on large, well-annotated remote sensing datasets. However, incomplete or inadequate ground truth can hinder these learning approaches. Many infrastructure networks have graph-level properties such as reachability to a source (like canals) or connectivity (roads) that can be leveraged to improve these existing ground truth. This paper develops a novel iterative framework IGraSS, combining a semantic segmentation module-incorporating RGB and additional modalities (NDWI, DEM)-with a graph-based ground-truth refinement module. The segmentation module processes satellite imagery patches, while the refinement module operates on the entire data viewing the infrastructure network as a graph. Experiments show that IGraSS reduces unreachable canal segments from around 18% to 3%, and training with refined ground truth significantly improves canal identification. IGraSS serves as a robust framework for both refining noisy ground truth and mapping canal networks from remote sensing imagery. We also demonstrate the effectiveness and generalizability of IGraSS using road networks as an example, applying a different graph-theoretic constraint to complete road networks.
Authors: Mateusz Michalkiewicz, Anekha Sokhal, Tadeusz Michalkiewicz, Piotr Pawlikowski, Mahsa Baktashmotlagh, Varun Jampani, Guha Balakrishnan
Abstract: Monocular 3D reconstruction methods and vision-language models (VLMs) demonstrate impressive results on standard benchmarks, yet their true understanding of geometric properties remains unclear. We introduce GIQ , a comprehensive benchmark specifically designed to evaluate the geometric reasoning capabilities of vision and vision-language foundation models. GIQ comprises synthetic and real-world images of 224 diverse polyhedra - including Platonic, Archimedean, Johnson, and Catalan solids, as well as stellations and compound shapes - covering varying levels of complexity and symmetry. Through systematic experiments involving monocular 3D reconstruction, 3D symmetry detection, mental rotation tests, and zero-shot shape classification tasks, we reveal significant shortcomings in current models. State-of-the-art reconstruction algorithms trained on extensive 3D datasets struggle to reconstruct even basic geometric forms accurately. While foundation models effectively detect specific 3D symmetry elements via linear probing, they falter significantly in tasks requiring detailed geometric differentiation, such as mental rotation. Moreover, advanced vision-language assistants exhibit remarkably low accuracy on complex polyhedra, systematically misinterpreting basic properties like face geometry, convexity, and compound structures. GIQ is publicly available, providing a structured platform to highlight and address critical gaps in geometric intelligence, facilitating future progress in robust, geometry-aware representation learning.
Authors: Guandong Li, Mengxia Ye
Abstract: Deep neural networks face several challenges in hyperspectral image classification, including high-dimensional data, sparse distribution of ground objects, and spectral redundancy, which often lead to classification overfitting and limited generalization capability. To more effectively extract and fuse spatial context with fine spectral information in hyperspectral image (HSI) classification, this paper proposes a novel network architecture called STNet. The core advantage of STNet stems from the dual innovative design of its Spatial-Spectral Transformer module: first, the fundamental explicit decoupling of spatial and spectral attention ensures targeted capture of key information in HSI; second, two functionally distinct gating mechanisms perform intelligent regulation at both the fusion level of attention flows (adaptive attention fusion gating) and the internal level of feature transformation (GFFN). This characteristic demonstrates superior feature extraction and fusion capabilities compared to traditional convolutional neural networks, while reducing overfitting risks in small-sample and high-noise scenarios. STNet enhances model representation capability without increasing network depth or width. The proposed method demonstrates superior performance on IN, UP, and KSC datasets, outperforming mainstream hyperspectral image classification approaches.
Authors: Shivang Chopra, Gabriela Sanchez-Rodriguez, Lingchao Mao, Andrew J Feola, Jing Li, Zsolt Kira
Abstract: Different medical imaging modalities capture diagnostic information at varying spatial resolutions, from coarse global patterns to fine-grained localized structures. However, most existing vision-language frameworks in the medical domain apply a uniform strategy for local feature extraction, overlooking the modality-specific demands. In this work, we present MedMoE, a modular and extensible vision-language processing framework that dynamically adapts visual representation based on the diagnostic context. MedMoE incorporates a Mixture-of-Experts (MoE) module conditioned on the report type, which routes multi-scale image features through specialized expert branches trained to capture modality-specific visual semantics. These experts operate over feature pyramids derived from a Swin Transformer backbone, enabling spatially adaptive attention to clinically relevant regions. This framework produces localized visual representations aligned with textual descriptions, without requiring modality-specific supervision at inference. Empirical results on diverse medical benchmarks demonstrate that MedMoE improves alignment and retrieval performance across imaging modalities, underscoring the value of modality-specialized visual representations in clinical vision-language systems.
Authors: Peter Gr\"onquist, Stepan Tulyakov, Dengxin Dai
Abstract: Achieving consistent color reproduction across multiple cameras is essential for seamless image fusion and Image Processing Pipeline (ISP) compatibility in modern devices, but it is a challenging task due to variations in sensors and optics. Existing raw-to-raw conversion methods face limitations such as poor adaptability to changing illumination, high computational costs, or impractical requirements such as simultaneous camera operation and overlapping fields-of-view. We introduce the Neural Physical Model (NPM), a lightweight, physically-informed approach that simulates raw images under specified illumination to estimate transformations between devices. The NPM effectively adapts to varying illumination conditions, can be initialized with physical measurements, and supports training with or without paired data. Experiments on public datasets like NUS and BeyondRGB demonstrate that NPM outperforms recent state-of-the-art methods, providing robust chromatic consistency across different sensors and optical systems.
Authors: Dieuwertje Alblas, Patryk Rygiel, Julian Suk, Kaj O. Kappe, Marieke Hofman, Christoph Brune, Kak Khee Yeung, Jelmer M. Wolterink
Abstract: Abdominal aortic aneurysms (AAAs) are progressive focal dilatations of the abdominal aorta. AAAs may rupture, with a survival rate of only 20\%. Current clinical guidelines recommend elective surgical repair when the maximum AAA diameter exceeds 55 mm in men or 50 mm in women. Patients that do not meet these criteria are periodically monitored, with surveillance intervals based on the maximum AAA diameter. However, this diameter does not take into account the complex relation between the 3D AAA shape and its growth, making standardized intervals potentially unfit. Personalized AAA growth predictions could improve monitoring strategies. We propose to use an SE(3)-symmetric transformer model to predict AAA growth directly on the vascular model surface enriched with local, multi-physical features. In contrast to other works which have parameterized the AAA shape, this representation preserves the vascular surface's anatomical structure and geometric fidelity. We train our model using a longitudinal dataset of 113 computed tomography angiography (CTA) scans of 24 AAA patients at irregularly sampled intervals. After training, our model predicts AAA growth to the next scan moment with a median diameter error of 1.18 mm. We further demonstrate our model's utility to identify whether a patient will become eligible for elective repair within two years (acc = 0.93). Finally, we evaluate our model's generalization on an external validation set consisting of 25 CTAs from 7 AAA patients from a different hospital. Our results show that local directional AAA growth prediction from the vascular surface is feasible and may contribute to personalized surveillance strategies.
Authors: Jiayi Song, Kaiyu Li, Xiangyong Cao, Deyu Meng
Abstract: Semantic segmentation in remote sensing images is crucial for various applications, yet its performance is heavily reliant on large-scale, high-quality pixel-wise annotations, which are notoriously expensive and time-consuming to acquire. Semi-supervised semantic segmentation (SSS) offers a promising alternative to mitigate this data dependency. However, existing SSS methods often struggle with the inherent distribution mismatch between limited labeled data and abundant unlabeled data, leading to suboptimal generalization. To alleviate this issue, we attempt to introduce the Vision Foundation Models (VFMs) pre-trained on vast and diverse datasets into the SSS task since VFMs possess robust generalization capabilities that can effectively bridge this distribution gap and provide strong semantic priors for SSS. Inspired by this, we introduce RS-MTDF (Multi-Teacher Distillation and Fusion), a novel framework that leverages the powerful semantic knowledge embedded in VFMs to guide semi-supervised learning in remote sensing. Specifically, RS-MTDF employs multiple frozen VFMs (e.g., DINOv2 and CLIP) as expert teachers, utilizing feature-level distillation to align student features with their robust representations. To further enhance discriminative power, the distilled knowledge is seamlessly fused into the student decoder. Extensive experiments on three challenging remote sensing datasets demonstrate that RS-MTDF consistently achieves state-of-the-art performance. Notably, our method outperforms existing approaches across various label ratios on LoveDA and secures the highest IoU in the majority of semantic categories. These results underscore the efficacy of multi-teacher VFM guidance in significantly enhancing both generalization and semantic understanding for remote sensing segmentation. Ablation studies further validate the contribution of each proposed module.
Authors: Keyi Liu, Weidong Yang, Ben Fei, Ying He
Abstract: Self-supervised learning (SSL) for point cloud pre-training has become a cornerstone for many 3D vision tasks, enabling effective learning from large-scale unannotated data. At the scene level, existing SSL methods often incorporate volume rendering into the pre-training framework, using RGB-D images as reconstruction signals to facilitate cross-modal learning. This strategy promotes alignment between 2D and 3D modalities and enables the model to benefit from rich visual cues in the RGB-D inputs. However, these approaches are limited by their reliance on implicit scene representations and high memory demands. Furthermore, since their reconstruction objectives are applied only in 2D space, they often fail to capture underlying 3D geometric structures. To address these challenges, we propose Gaussian2Scene, a novel scene-level SSL framework that leverages the efficiency and explicit nature of 3D Gaussian Splatting (3DGS) for pre-training. The use of 3DGS not only alleviates the computational burden associated with volume rendering but also supports direct 3D scene reconstruction, thereby enhancing the geometric understanding of the backbone network. Our approach follows a progressive two-stage training strategy. In the first stage, a dual-branch masked autoencoder learns both 2D and 3D scene representations. In the second stage, we initialize training with reconstructed point clouds and further supervise learning using the geometric locations of Gaussian primitives and rendered RGB images. This process reinforces both geometric and cross-modal learning. We demonstrate the effectiveness of Gaussian2Scene across several downstream 3D object detection tasks, showing consistent improvements over existing pre-training methods.
Authors: Shuyi Zhang, Xiaoshuai Hao, Yingbo Tang, Lingfeng Zhang, Pengwei Wang, Zhongyuan Wang, Hongxuan Ma, Shanghang Zhang
Abstract: Video content comprehension is essential for various applications, ranging from video analysis to interactive systems. Despite advancements in large-scale vision-language models (VLMs), these models often struggle to capture the nuanced, spatiotemporal details essential for thorough video analysis. To address this gap, we introduce Video-CoT, a groundbreaking dataset designed to enhance spatiotemporal understanding using Chain-of-Thought (CoT) methodologies. Video-CoT contains 192,000 fine-grained spa-tiotemporal question-answer pairs and 23,000 high-quality CoT-annotated samples, providing a solid foundation for evaluating spatiotemporal understanding in video comprehension. Additionally, we provide a comprehensive benchmark for assessing these tasks, with each task featuring 750 images and tailored evaluation metrics. Our extensive experiments reveal that current VLMs face significant challenges in achieving satisfactory performance, high-lighting the difficulties of effective spatiotemporal understanding. Overall, the Video-CoT dataset and benchmark open new avenues for research in multimedia understanding and support future innovations in intelligent systems requiring advanced video analysis capabilities. By making these resources publicly available, we aim to encourage further exploration in this critical area. Project website:https://video-cot.github.io/ .
Authors: Jingguo Qu, Xinyang Han, Tonghuan Xiao, Jia Ai, Juan Wu, Tong Zhao, Jing Qin, Ann Dorothy King, Winnie Chiu-Wing Chu, Jing Cai, Michael Tin-Cheung Ying
Abstract: Medical ultrasonography is an essential imaging technique for examining superficial organs and tissues, including lymph nodes, breast, and thyroid. It employs high-frequency ultrasound waves to generate detailed images of the internal structures of the human body. However, manually contouring regions of interest in these images is a labor-intensive task that demands expertise and often results in inconsistent interpretations among individuals. Vision-language foundation models, which have excelled in various computer vision applications, present new opportunities for enhancing ultrasound image analysis. Yet, their performance is hindered by the significant differences between natural and medical imaging domains. This research seeks to overcome these challenges by developing domain adaptation methods for vision-language foundation models. In this study, we explore the fine-tuning pipeline for vision-language foundation models by utilizing large language model as text refiner with special-designed adaptation strategies and task-driven heads. Our approach has been extensively evaluated on six ultrasound datasets and two tasks: segmentation and classification. The experimental results show that our method can effectively improve the performance of vision-language foundation models for ultrasound image analysis, and outperform the existing state-of-the-art vision-language and pure foundation models. The source code of this study is available at https://github.com/jinggqu/NextGen-UIA.
Authors: Jos\'e Morano, Botond Fazekas, Emese S\"ukei, Ronald Fecso, Taha Emre, Markus Gumpinger, Georg Faustmann, Marzieh Oghbaie, Ursula Schmidt-Erfurth, Hrvoje Bogunovi\'c
Abstract: Artificial intelligence (AI) has become a fundamental tool for assisting clinicians in analyzing ophthalmic images, such as optical coherence tomography (OCT). However, developing AI models often requires extensive annotation, and existing models tend to underperform on independent, unseen data. Foundation models (FMs), large AI models trained on vast unlabeled datasets, have shown promise in overcoming these challenges. Nonetheless, available FMs for ophthalmology lack extensive validation, especially for segmentation tasks, and focus on a single imaging modality. In this context, we propose MIRAGE, a novel multimodal FM for the analysis of OCT and scanning laser ophthalmoscopy (SLO) images. Additionally, we propose a new evaluation benchmark with OCT/SLO classification and segmentation tasks. The comparison with general and specialized FMs and segmentation methods shows the superiority of MIRAGE in both types of tasks, highlighting its suitability as a basis for the development of robust AI systems for retinal OCT image analysis. Both MIRAGE and the evaluation benchmark are publicly available: https://github.com/j-morano/MIRAGE.
Authors: Jiajun Li (University of Electronic Science and Technology of China, Shanghai Jiaotong University), Yue Ma (The Hong Kong University of Science and Technology), Xinyu Zhang (University of Electronic Science and Technology of China), Qingyan Wei (Central South University), Songhua Liu (National University of Singapore, Shanghai Jiaotong University), Linfeng Zhang (Shanghai Jiaotong University)
Abstract: Recent studies on Visual Autoregressive (VAR) models have highlighted that high-frequency components, or later steps, in the generation process contribute disproportionately to inference latency. However, the underlying computational redundancy involved in these steps has yet to be thoroughly investigated. In this paper, we conduct an in-depth analysis of the VAR inference process and identify two primary sources of inefficiency: step redundancy and unconditional branch redundancy. To address step redundancy, we propose an automatic step-skipping strategy that selectively omits unnecessary generation steps to improve efficiency. For unconditional branch redundancy, we observe that the information gap between the conditional and unconditional branches is minimal. Leveraging this insight, we introduce unconditional branch replacement, a technique that bypasses the unconditional branch to reduce computational cost. Notably, we observe that the effectiveness of acceleration strategies varies significantly across different samples. Motivated by this, we propose SkipVAR, a sample-adaptive framework that leverages frequency information to dynamically select the most suitable acceleration strategy for each instance. To evaluate the role of high-frequency information, we introduce high-variation benchmark datasets that test model sensitivity to fine details. Extensive experiments show SkipVAR achieves over 0.88 average SSIM with up to 1.81x overall acceleration and 2.62x speedup on the GenEval benchmark, maintaining model quality. These results confirm the effectiveness of frequency-aware, training-free adaptive acceleration for scalable autoregressive image generation. Our code is available at https://github.com/fakerone-li/SkipVAR and has been publicly released.
Authors: Daniel Shao, Richard J. Chen, Andrew H. Song, Joel Runevic, Ming Y. Lu, Tong Ding, Faisal Mahmood
Abstract: Multiple Instance Learning (MIL) is a cornerstone approach in computational pathology (CPath) for generating clinically meaningful slide-level embeddings from gigapixel tissue images. However, MIL often struggles with small, weakly supervised clinical datasets. In contrast to fields such as NLP and conventional computer vision, where transfer learning is widely used to address data scarcity, the transferability of MIL models remains poorly understood. In this study, we systematically evaluate the transfer learning capabilities of pretrained MIL models by assessing 11 models across 21 pretraining tasks for morphological and molecular subtype prediction. Our results show that pretrained MIL models, even when trained on different organs than the target task, consistently outperform models trained from scratch. Moreover, pretraining on pancancer datasets enables strong generalization across organs and tasks, outperforming slide foundation models while using substantially less pretraining data. These findings highlight the robust adaptability of MIL models and demonstrate the benefits of leveraging transfer learning to boost performance in CPath. Lastly, we provide a resource which standardizes the implementation of MIL models and collection of pretrained model weights on popular CPath tasks, available at https://github.com/mahmoodlab/MIL-Lab
Authors: Juhyung Park, Dongwon Park, Sooyeon Ji, Hyeong-Geol Shin, Se Young Chun, Jongho Lee
Abstract: Denoising of magnetic resonance images is beneficial in improving the quality of low signal-to-noise ratio images. Recently, denoising using deep neural networks has demonstrated promising results. Most of these networks, however, utilize supervised learning, which requires large training images of noise-corrupted and clean image pairs. Obtaining training images, particularly clean images, is expensive and time-consuming. Hence, methods such as Noise2Noise (N2N) that require only pairs of noise-corrupted images have been developed to reduce the burden of obtaining training datasets. In this study, we propose a new self-supervised denoising method, Coil2Coil (C2C), that does not require the acquisition of clean images or paired noise-corrupted images for training. Instead, the method utilizes multichannel data from phased-array coils to generate training images. First, it divides and combines multichannel coil images into two images, one for input and the other for label. Then, they are processed to impose noise independence and sensitivity normalization such that they can be used for the training images of N2N. For inference, the method inputs a coil-combined image (e.g., DICOM image), enabling a wide application of the method. When evaluated using synthetic noise-added images, C2C shows the best performance against several self-supervised methods, reporting comparable outcomes to supervised methods. When testing the DICOM images, C2C successfully denoised real noise without showing structure-dependent residuals in the error maps. Because of the significant advantage of not requiring additional scans for clean or paired images, the method can be easily utilized for various clinical applications.
Authors: Ahmad Rahimi, Po-Chien Luan, Yuejiang Liu, Frano Raji\v{c}, Alexandre Alahi
Abstract: Modeling spatial-temporal interactions among neighboring agents is at the heart of multi-agent problems such as motion forecasting and crowd navigation. Despite notable progress, it remains unclear to which extent modern representations can capture the causal relationships behind agent interactions. In this work, we take an in-depth look at the causal awareness of these representations, from computational formalism to real-world practice. First, we cast doubt on the notion of non-causal robustness studied in the recent CausalAgents benchmark. We show that recent representations are already partially resilient to perturbations of non-causal agents, and yet modeling indirect causal effects involving mediator agents remains challenging. To address this challenge, we introduce a metric learning approach that regularizes latent representations with causal annotations. Our controlled experiments show that this approach not only leads to higher degrees of causal awareness but also yields stronger out-of-distribution robustness. To further operationalize it in practice, we propose a sim-to-real causal transfer method via cross-domain multi-task learning. Experiments on pedestrian datasets show that our method can substantially boost generalization, even in the absence of real-world causal annotations. We hope our work provides a new perspective on the challenges and pathways towards causally-aware representations of multi-agent interactions. Our code is available at https://github.com/vita-epfl/CausalSim2Real.
Authors: Marien Renaud, Jean Prost, Arthur Leclaire, Nicolas Papadakis
Abstract: Plug-and-Play (PnP) algorithms are a class of iterative algorithms that address image inverse problems by combining a physical model and a deep neural network for regularization. Even if they produce impressive image restoration results, these algorithms rely on a non-standard use of a denoiser on images that are less and less noisy along the iterations, which contrasts with recent algorithms based on Diffusion Models (DM), where the denoiser is applied only on re-noised images. We propose a new PnP framework, called Stochastic deNOising REgularization (SNORE), which applies the denoiser only on images with noise of the adequate level. It is based on an explicit stochastic regularization, which leads to a stochastic gradient descent algorithm to solve ill-posed inverse problems. A convergence analysis of this algorithm and its annealing extension is provided. Experimentally, we prove that SNORE is competitive with respect to state-of-the-art methods on deblurring and inpainting tasks, both quantitatively and qualitatively.
Authors: Divyansh Singhvi, Diganta Misra, Andrej Erkelens, Raghav Jain, Isabel Papadimitriou, Naomi Saphra
Abstract: Language is an intricately structured system, and a key goal of NLP interpretability is to provide methodological insights for understanding how language models represent this structure internally. In this paper, we use Shapley Taylor interaction indices (STII) in order to examine how language and speech models internally relate and structure their inputs. Pairwise Shapley interactions measure how much two inputs work together to influence model outputs beyond if we linearly added their independent influences, providing a view into how models encode structural interactions between inputs. We relate the interaction patterns in models to three underlying linguistic structures: syntactic structure, non-compositional semantics, and phonetic coarticulation. We find that autoregressive text models encode interactions that correlate with the syntactic proximity of inputs, and that both autoregressive and masked models encode nonlinear interactions in idiomatic phrases with non-compositional semantics. Our speech results show that inputs are more entangled for pairs where a neighboring consonant is likely to influence a vowel or approximant, showing that models encode the phonetic interaction needed for extracting discrete phonemic representations.
Authors: Mahrokh Najaf, Gregory Ongie
Abstract: Implicit neural representations (INRs) have emerged as a powerful tool for solving inverse problems in computer vision and computational imaging. INRs represent images as continuous domain functions realized by a neural network taking spatial coordinates as inputs. However, unlike traditional pixel representations, little is known about the sample complexity of estimating images using INRs in the context of linear inverse problems. Towards this end, we study the sampling requirements for recovery of a continuous domain image from its low-pass Fourier coefficients by fitting a single hidden-layer INR with ReLU activation and a Fourier features layer using a generalized form of weight decay regularization. Our key insight is to relate minimizers of this non-convex parameter space optimization problem to minimizers of a convex penalty defined over an infinite-dimensional space of measures. We identify a sufficient number of samples for which an image realized by a width-1 INR is exactly recoverable by solving the INR training problem, and give a conjecture for the general width-$W$ case. To validate our theory, we empirically assess the probability of achieving exact recovery of images realized by low-width single hidden-layer INRs, and illustrate the performance of INR on super-resolution recovery of more realistic continuous domain phantom images.
Authors: Melvin Wong, Jiao Liu, Thiago Rios, Stefan Menzel, Yew Soon Ong
Abstract: In this paper, we introduce LLM-driven MultiTask Evolutionary Algorithm (LLM2TEA), the first agentic AI designer within a generative evolutionary multitasking (GEM) framework that promotes the crossover and synergy of designs from multiple domains, leading to innovative solutions that transcend individual disciplines. Of particular interest is the discovery of objects that are not only innovative but also conform to the physical specifications of the real world in science and engineering. LLM2TEA comprises a large language model to initialize a population of genotypes (defined by text prompts) describing the objects of interest, a text-to-3D generative model to produce phenotypes from these prompts, a classifier to interpret the semantic representations of the objects, and a physics simulation model to assess their physical properties. We propose several novel LLM-based multitask evolutionary operators to guide the search toward the discovery of high-performing practical objects. Experimental results in conceptual design optimization validate the effectiveness of LLM2TEA, revealing from 97\% to 174\% improvement in the diversity of innovative objects compared to the present text-to-3D generative model baseline. In addition, more than 73\% of the generated designs have better physical performance than the top 1\% percentile of the designs generated in the baseline. Moreover, LLM2TEA generates designs that are not only aesthetically creative but also functional in real-world applications. Several of these designs have been successfully 3D-printed, emphasizing the proposed approach's capacity to transform AI-generated outputs into tangible physical objects. The designs produced by LLM2TEA meets practical requirements while showcasing creative and innovative features, underscoring its potential applications in complex design optimization and discovery.
Authors: Changde Du, Kaicheng Fu, Bincheng Wen, Yi Sun, Jie Peng, Wei Wei, Ying Gao, Shengpei Wang, Chuncheng Zhang, Jinpeng Li, Shuang Qiu, Le Chang, Huiguang He
Abstract: Understanding how humans conceptualize and categorize natural objects offers critical insights into perception and cognition. With the advent of Large Language Models (LLMs), a key question arises: can these models develop human-like object representations from linguistic and multimodal data? In this study, we combined behavioral and neuroimaging analyses to explore the relationship between object concept representations in LLMs and human cognition. We collected 4.7 million triplet judgments from LLMs and Multimodal LLMs (MLLMs) to derive low-dimensional embeddings that capture the similarity structure of 1,854 natural objects. The resulting 66-dimensional embeddings were stable, predictive, and exhibited semantic clustering similar to human mental representations. Remarkably, the dimensions underlying these embeddings were interpretable, suggesting that LLMs and MLLMs develop human-like conceptual representations of objects. Further analysis showed strong alignment between model embeddings and neural activity patterns in brain regions such as EBA, PPA, RSC, and FFA. This provides compelling evidence that the object representations in LLMs, while not identical to human ones, share fundamental similarities that reflect key aspects of human conceptual knowledge. Our findings advance the understanding of machine intelligence and inform the development of more human-like artificial cognitive systems.
Authors: Yanbing Chen, Tianze Tang, Taehyo Kim, Hai Shu
Abstract: Gliomas are among the most common malignant brain tumors and are characterized by considerable heterogeneity, which complicates accurate detection and segmentation. Multi-modal MRI is the clinical standard for glioma imaging, but variability across modalities and high computational complexity hinder effective automated segmentation. In this paper, we propose UKAN-EP, a novel 3D extension of the original 2D U-KAN model for multi-modal MRI brain tumor segmentation. While U-KAN integrates Kolmogorov-Arnold Network (KAN) layers into a U-Net backbone, UKAN-EP further incorporates Efficient Channel Attention (ECA) and Pyramid Feature Aggregation (PFA) modules to enhance inter-modality feature fusion and multi-scale feature representation. We also introduce a dynamic loss weighting strategy that adaptively balances the Cross-Entropy and Dice losses during training. We evaluate UKAN-EP on the 2024 BraTS-GLI dataset and compare it against strong baselines including U-Net, Attention U-Net, and Swin UNETR. Results show that UKAN-EP achieves superior segmentation performance while requiring substantially fewer computational resources. An extensive ablation study further demonstrates the effectiveness of ECA and PFA, as well as the limited utility of self-attention and spatial attention alternatives. Code is available at https://github.com/TianzeTang0504/UKAN-EP.
Authors: Kirsten W. H. Maas, Danny Ruijters, Anna Vilanova, Nicola Pezzotti
Abstract: Dynamic three-dimensional (4D) reconstruction from two-dimensional X-ray coronary angiography (CA) remains a significant clinical problem. Existing CA reconstruction methods often require extensive user interaction or large training datasets. Recently, Neural Radiance Field (NeRF) has successfully reconstructed high-fidelity scenes in natural and medical contexts without these requirements. However, challenges such as sparse-views, intra-scan motion, and complex vessel morphology hinder its direct application to CA data. We introduce NeRF-CA, a first step toward a fully automatic 4D CA reconstruction that achieves reconstructions from sparse coronary angiograms. To the best of our knowledge, we are the first to address the challenges of sparse-views and cardiac motion by decoupling the scene into the moving coronary artery and the static background, effectively translating the problem of motion into a strength. NeRF-CA serves as a first stepping stone for solving the 4D CA reconstruction problem, achieving adequate 4D reconstructions from as few as four angiograms, as required by clinical practice, while significantly outperforming state-of-the-art sparse-view X-ray NeRF. We validate our approach quantitatively and qualitatively using representative 4D phantom datasets and ablation studies. To accelerate research in this domain, we made our codebase public: https://github.com/kirstenmaas/NeRF-CA.
Authors: Ning Zhang, Timothy Shea, Arto Nurmikko
Abstract: Tracking and acquiring simultaneous optical images of randomly moving targets obscured by scattering media remains a challenging problem of importance to many applications that require precise object localization and identification. In this work we develop an end-to-end neuromorphic optical engineering and computational approach to demonstrate how to track and image normally invisible objects by combining an event detecting camera with a multistage neuromorphic deep learning strategy. Photons emerging from dense scattering media are detected by the event camera and converted to pixel-wise asynchronized spike trains - a first step in isolating object-specific information from the dominant uninformative background. Spiking data is fed into a deep spiking neural network (SNN) engine where object tracking and image reconstruction are performed by two separate yet interconnected modules running in parallel in discrete time steps over the event duration. Through benchtop experiments we demonstrate tracking and imaging randomly moving objects in dense turbid media as well as image reconstruction of spatially stationary but optically dynamic objects. Standardized character sets serve as representative proxies for geometrically complex objects, underscoring the method's generality. The results highlight the advantages of a fully neuromorphic approach in meeting a major imaging technology with high computational efficiency and low power consumption.
Authors: Hou In Ivan Tam, Hou In Derek Pun, Austin T. Wang, Angel X. Chang, Manolis Savva
Abstract: Despite recent advances in text-conditioned 3D indoor scene generation, there remain gaps in the evaluation of these methods. Existing metrics primarily assess the realism of generated scenes by comparing them to a set of ground-truth scenes, often overlooking alignment with the input text - a critical factor in determining how effectively a method meets user requirements. We present SceneEval, an evaluation framework designed to address this limitation. SceneEval includes metrics for both explicit user requirements, such as the presence of specific objects and their attributes described in the input text, and implicit expectations, like the absence of object collisions, providing a comprehensive assessment of scene quality. To facilitate evaluation, we introduce SceneEval-500, a dataset of scene descriptions with annotated ground-truth scene properties. We evaluate recent scene generation methods using SceneEval and demonstrate its ability to provide detailed assessments of the generated scenes, highlighting strengths and areas for improvement across multiple dimensions. Our results show that current methods struggle at generating scenes that meet user requirements, underscoring the need for further research in this direction.
Authors: Ezzeldin Shereen, Dan Ristea, Shae McFadden, Burak Hasircioglu, Vasilios Mavroudis, Chris Hicks
Abstract: Multi-modal retrieval augmented generation (M-RAG) is instrumental for inhibiting hallucinations in large multi-modal models (LMMs) through the use of a factual knowledge base (KB). However, M-RAG introduces new attack vectors for adversaries that aim to disrupt the system by injecting malicious entries into the KB. In this paper, we present the first poisoning attack against M-RAG targeting visual document retrieval applications where the KB contains images of document pages. We propose two attacks, each of which require injecting only a single adversarial image into the KB. Firstly, we propose a universal attack that, for any potential user query, influences the response to cause a denial-of-service (DoS) in the M-RAG system. Secondly, we present a targeted attack against one or a group of user queries, with the goal of spreading targeted misinformation. For both attacks, we use a multi-objective gradient-based adversarial approach to craft the injected image while optimizing for both retrieval and generation. We evaluate our attacks against several visual document retrieval datasets, a diverse set of state-of-the-art retrievers (embedding models) and generators (LMMs), demonstrating the attack effectiveness in both the universal and targeted settings. We additionally present results including commonly used defenses, various attack hyper-parameter settings, ablations, and attack transferability.
Authors: Blake VanBerlo, Alexander Wong, Jesse Hoey, Robert Arntfield
Abstract: Data augmentation is a central component of joint embedding self-supervised learning (SSL). Approaches that work for natural images may not always be effective in medical imaging tasks. This study systematically investigated the impact of data augmentation and preprocessing strategies in SSL for lung ultrasound. Three data augmentation pipelines were assessed: (1) a baseline pipeline commonly used across imaging domains, (2) a novel semantic-preserving pipeline designed for ultrasound, and (3) a distilled set of the most effective transformations from both pipelines. Pretrained models were evaluated on multiple classification tasks: B-line detection, pleural effusion detection, and COVID-19 classification. Experiments revealed that semantics-preserving data augmentation resulted in the greatest performance for COVID-19 classification - a diagnostic task requiring global image context. Cropping-based methods yielded the greatest performance on the B-line and pleural effusion object classification tasks, which require strong local pattern recognition. Lastly, semantics-preserving ultrasound image preprocessing resulted in increased downstream performance for multiple tasks. Guidance regarding data augmentation and preprocessing strategies was synthesized for practitioners working with SSL in ultrasound.
Authors: Arash Torabi Goodarzi, Roman Kochnev, Waleed Khalid, Furui Qin, Tolgay Atinc Uzun, Yashkumar Sanjaybhai Dhameliya, Yash Kanubhai Kathiriya, Zofia Antonina Bentyn, Dmitry Ignatov, Radu Timofte
Abstract: Neural networks are fundamental in artificial intelligence, driving progress in computer vision and natural language processing. High-quality datasets are crucial for their development, and there is growing interest in datasets composed of neural networks themselves to support benchmarking, automated machine learning (AutoML), and model analysis. We introduce LEMUR, an open source dataset of neural network models with well-structured code for diverse architectures across tasks such as object detection, image classification, segmentation, and natural language processing. LEMUR is primarily designed to provide a rich source of structured model representations and associated performance data, enabling the fine-tuning of large language models for AutoML applications. Leveraging Python and PyTorch, LEMUR enables seamless extension to new datasets and models while maintaining consistency. It integrates an Optuna-powered framework for evaluation, hyperparameter optimization, statistical analysis, and graphical insights. LEMUR VR extension enables the seamless deployment of models in virtual reality, optimizing their performance on resource-constrained devices. Providing tools for model evaluation, preprocessing, and database management, LEMUR supports researchers and practitioners in developing, testing, and analyzing neural networks. It offers an API that delivers comprehensive information about neural network models and their complete performance statistics with a single request, which can be used in experiments with code-generating large language models. The LEMUR and its plugins are accessible as open source projects under the MIT license at https://github.com/ABrain-One/nn-dataset, https://github.com/ABrain-One/nn-plots and https://github.com/ABrain-One/nn-vr.
URLs: https://github.com/ABrain-One/nn-dataset,, https://github.com/ABrain-One/nn-plots, https://github.com/ABrain-One/nn-vr.
Authors: Marcus J. Vroemen, Yuqian Chen, Yui Lo, Tengfei Xue, Weidong Cai, Fan Zhang, Josien P. W. Pluim, Lauren J. O'Donnell
Abstract: Diffusion MRI (dMRI) tractography enables in vivo mapping of brain structural connections, but traditional connectome generation is time-consuming and requires gray matter parcellation, posing challenges for large-scale studies. We introduce DeepMultiConnectome, a deep-learning model that predicts structural connectomes directly from tractography, bypassing the need for gray matter parcellation while supporting multiple parcellation schemes. Using a point-cloud-based neural network with multi-task learning, the model classifies streamlines according to their connected regions across two parcellation schemes, sharing a learned representation. We train and validate DeepMultiConnectome on tractography from the Human Connectome Project Young Adult dataset ($n = 1000$), labeled with an 84 and 164 region gray matter parcellation scheme. DeepMultiConnectome predicts multiple structural connectomes from a whole-brain tractogram containing 3 million streamlines in approximately 40 seconds. DeepMultiConnectome is evaluated by comparing predicted connectomes with traditional connectomes generated using the conventional method of labeling streamlines using a gray matter parcellation. The predicted connectomes are highly correlated with traditionally generated connectomes ($r = 0.992$ for an 84-region scheme; $r = 0.986$ for a 164-region scheme) and largely preserve network properties. A test-retest analysis of DeepMultiConnectome demonstrates reproducibility comparable to traditionally generated connectomes. The predicted connectomes perform similarly to traditionally generated connectomes in predicting age and cognitive function. Overall, DeepMultiConnectome provides a scalable, fast model for generating subject-specific connectomes across multiple parcellation schemes.
Authors: LASA Team, Weiwen Xu, Hou Pong Chan, Long Li, Mahani Aljunied, Ruifeng Yuan, Jianyu Wang, Chenghao Xiao, Guizhen Chen, Chaoqun Liu, Zhaodonghui Li, Yu Sun, Junao Shen, Chaojun Wang, Jie Tan, Deli Zhao, Tingyang Xu, Hao Zhang, Yu Rong
Abstract: Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in understanding common visual elements, largely due to their large-scale datasets and advanced training strategies. However, their effectiveness in medical applications remains limited due to the inherent discrepancies between data and tasks in medical scenarios and those in the general domain. Concretely, existing medical MLLMs face the following critical limitations: (1) limited coverage of medical knowledge beyond imaging, (2) heightened susceptibility to hallucinations due to suboptimal data curation processes, (3) lack of reasoning capabilities tailored for complex medical scenarios. To address these challenges, we first propose a comprehensive data curation procedure that (1) efficiently acquires rich medical knowledge data not only from medical imaging but also from extensive medical texts and general-domain data; and (2) synthesizes accurate medical captions, visual question answering (VQA), and reasoning samples. As a result, we build a multimodal dataset enriched with extensive medical knowledge. Building on the curated data, we introduce our medical-specialized MLLM: Lingshu. Lingshu undergoes multi-stage training to embed medical expertise and enhance its task-solving capabilities progressively. Besides, we preliminarily explore the potential of applying reinforcement learning with verifiable rewards paradigm to enhance Lingshu's medical reasoning ability. Additionally, we develop MedEvalKit, a unified evaluation framework that consolidates leading multimodal and textual medical benchmarks for standardized, fair, and efficient model assessment. We evaluate the performance of Lingshu on three fundamental medical tasks, multimodal QA, text-based QA, and medical report generation. The results show that Lingshu consistently outperforms the existing open-source multimodal models on most tasks ...
Authors: Philip R. Liu, Sparsh Bansal, Jimmy Dinh, Aditya Pawar, Ramani Satishkumar, Shail Desai, Neeraj Gupta, Xin Wang, Shu Hu
Abstract: The integration of deep learning-based glaucoma detection with large language models (LLMs) presents an automated strategy to mitigate ophthalmologist shortages and improve clinical reporting efficiency. However, applying general LLMs to medical imaging remains challenging due to hallucinations, limited interpretability, and insufficient domain-specific medical knowledge, which can potentially reduce clinical accuracy. Although recent approaches combining imaging models with LLM reasoning have improved reporting, they typically rely on a single generalist agent, restricting their capacity to emulate the diverse and complex reasoning found in multidisciplinary medical teams. To address these limitations, we propose MedChat, a multi-agent diagnostic framework and platform that combines specialized vision models with multiple role-specific LLM agents, all coordinated by a director agent. This design enhances reliability, reduces hallucination risk, and enables interactive diagnostic reporting through an interface tailored for clinical review and educational use. Code available at https://github.com/Purdue-M2/MedChat.
Authors: Chenxi Liu, Tianyi Xiong, Ruibo Chen, Yihan Wu, Junfeng Guo, Tianyi Zhou, Heng Huang
Abstract: The task adaptation and alignment of Large Multimodal Models (LMMs) have been significantly advanced by instruction tuning and further strengthened by recent preference optimization. Yet, most LMMs still suffer from severe modality imbalance during reasoning, i.e., outweighing language prior biases over visual inputs, which bottlenecks their generalization to downstream tasks and causes hallucinations. However, existing preference optimization approaches for LMMs do not focus on restraining the internal biases of their Large Language Model (LLM) backbones when curating the training data. Moreover, they heavily rely on offline data and lack the capacity to explore diverse responses adaptive to dynamic distributional shifts during training. Meanwhile, Group Relative Policy Optimization (GRPO), a recent method using online-generated data and verified rewards to improve reasoning capabilities, remains largely underexplored in LMM alignment. In this paper, we propose a novel preference learning framework, Modality-Balancing Preference Optimization (MBPO), to address the modality imbalance in LMMs. MBPO constructs a more effective offline preference dataset by generating hard negatives, i.e., rejected responses misled by LLM biases due to limited usage of visual information, through adversarial perturbation of input images. Moreover, MBPO leverages the easy-to-verify nature of close-ended tasks to generate online responses with verified rewards. GRPO is then employed to train the model with offline-online hybrid data. Extensive experiments demonstrate that MBPO can enhance LMM performance on challenging vision-language tasks and effectively reduce hallucinations.
Authors: Ruiran Su, Jiasheng Si, Zhijiang Guo, Janet B. Pierrehumbert
Abstract: Scientific fact-checking has mostly focused on text and tables, overlooking scientific charts, which are key for presenting quantitative evidence and statistical reasoning. We introduce ClimateViz, the first large-scale benchmark for scientific fact-checking using expert-curated scientific charts. ClimateViz contains 49,862 claims linked to 2,896 visualizations, each labeled as support, refute, or not enough information. To improve interpretability, each example includes structured knowledge graph explanations covering trends, comparisons, and causal relations. We evaluate state-of-the-art multimodal language models, including both proprietary and open-source systems, in zero-shot and few-shot settings. Results show that current models struggle with chart-based reasoning: even the best systems, such as Gemini 2.5 and InternVL 2.5, reach only 76.2 to 77.8 percent accuracy in label-only settings, far below human performance (89.3 and 92.7 percent). Explanation-augmented outputs improve performance in some models. We released our dataset and code alongside the paper.
Authors: Julia Guerrero-Viu, Michael Fischer, Iliyan Georgiev, Elena Garces, Diego Gutierrez, Belen Masia, Valentin Deschaintre
Abstract: Selection is the first step in many image editing processes, enabling faster and simpler modifications of all pixels sharing a common modality. In this work, we present a method for material selection in images, robust to lighting and reflectance variations, which can be used for downstream editing tasks. We rely on vision transformer (ViT) models and leverage their features for selection, proposing a multi-resolution processing strategy that yields finer and more stable selection results than prior methods. Furthermore, we enable selection at two levels: texture and subtexture, leveraging a new two-level material selection (DuMaS) dataset which includes dense annotations for over 800,000 synthetic images, both on the texture and subtexture levels.