Authors: Zongyou Yang, Jonathan Loo
Abstract: Recently, a significant improvement in the accuracy of 3D human pose estimation has been achieved by combining convolutional neural networks (CNNs) with pyramid grid alignment feedback loops. Additionally, innovative breakthroughs have been made in the field of computer vision through the adoption of Transformer-based temporal analysis architectures. Given these advancements, this study aims to deeply optimize and improve the existing Pymaf network architecture. The main innovations of this paper include: (1) Introducing a Transformer feature extraction network layer based on self-attention mechanisms to enhance the capture of low-level features; (2) Enhancing the understanding and capture of temporal signals in video sequences through feature temporal fusion techniques; (3) Implementing spatial pyramid structures to achieve multi-scale feature fusion, effectively balancing feature representations differences across different scales. The new PyCAT4 model obtained in this study is validated through experiments on the COCO and 3DPW datasets. The results demonstrate that the proposed improvement strategies significantly enhance the network's detection capability in human pose estimation, further advancing the development of human pose estimation technology.
Authors: Tongchun Zuo, Zaiyu Huang, Shuliang Ning, Ente Lin, Chao Liang, Zerong Zheng, Jianwen Jiang, Yuan Zhang, Mingyuan Gao, Xin Dong
Abstract: Video virtual try-on (VVT) technology has garnered considerable academic interest owing to its promising applications in e-commerce advertising and entertainment. However, most existing end-to-end methods rely heavily on scarce paired garment-centric datasets and fail to effectively leverage priors of advanced visual models and test-time inputs, making it challenging to accurately preserve fine-grained garment details and maintain temporal consistency in unconstrained scenarios. To address these challenges, we propose DreamVVT, a carefully designed two-stage framework built upon Diffusion Transformers (DiTs), which is inherently capable of leveraging diverse unpaired human-centric data to enhance adaptability in real-world scenarios. To further leverage prior knowledge from pretrained models and test-time inputs, in the first stage, we sample representative frames from the input video and utilize a multi-frame try-on model integrated with a vision-language model (VLM), to synthesize high-fidelity and semantically consistent keyframe try-on images. These images serve as complementary appearance guidance for subsequent video generation. \textbf{In the second stage}, skeleton maps together with fine-grained motion and appearance descriptions are extracted from the input content, and these along with the keyframe try-on images are then fed into a pretrained video generation model enhanced with LoRA adapters. This ensures long-term temporal coherence for unseen regions and enables highly plausible dynamic motions. Extensive quantitative and qualitative experiments demonstrate that DreamVVT surpasses existing methods in preserving detailed garment content and temporal stability in real-world scenarios. Our project page https://virtu-lab.github.io/
Authors: Adam Colton
Abstract: In the standard image joint embedding predictive architecture (IJEPA), features at the output of the teacher encoder are layer normalized (LN) before serving as a distillation target for the student encoder and predictor. We propose that this feature normalization disrupts the natural energy hierarchy of visual tokens, where high-energy tokens (those with larger L2 norms) encode semantically important image regions. LN forces all features to have identical L2 norms, effectively equalizing their energies and preventing the model from prioritizing semantically rich regions. We find that IJEPA models trained with feature LN exhibit loss maps with significant checkerboard-like artifacts. We propose that feature LN be replaced with a DynTanh activation as the latter better preserves token energies and allows high-energy tokens to greater contribute to the prediction loss. We show that IJEPA trained with feature DynTanh exhibits a longer-tailed loss distribution and fixes the checkerboard artifacts in the loss map. Our empirical results show that our simple modification improves ImageNet linear probe accuracy from 38% to 42.7% for ViT-Small and reduces RMSE by 0.08 on NYU Depth V2 monocular depth estimation. These results suggest that preserving natural token energies is crucial for effective self-supervised visual representation learning.
Authors: Miko{\l}aj Zieli\'nski, Krzysztof Byrski, Tomasz Szczepanik, Przemys{\l}aw Spurek
Abstract: Neural Radiance Fields (NeRF) and Gaussian Splatting (GS) have recently transformed 3D scene representation and rendering. NeRF achieves high-fidelity novel view synthesis by learning volumetric representations through neural networks, but its implicit encoding makes editing and physical interaction challenging. In contrast, GS represents scenes as explicit collections of Gaussian primitives, enabling real-time rendering, faster training, and more intuitive manipulation. This explicit structure has made GS particularly well-suited for interactive editing and integration with physics-based simulation. In this paper, we introduce GENIE (Gaussian Encoding for Neural Radiance Fields Interactive Editing), a hybrid model that combines the photorealistic rendering quality of NeRF with the editable and structured representation of GS. Instead of using spherical harmonics for appearance modeling, we assign each Gaussian a trainable feature embedding. These embeddings are used to condition a NeRF network based on the k nearest Gaussians to each query point. To make this conditioning efficient, we introduce Ray-Traced Gaussian Proximity Search (RT-GPS), a fast nearest Gaussian search based on a modified ray-tracing pipeline. We also integrate a multi-resolution hash grid to initialize and update Gaussian features. Together, these components enable real-time, locality-aware editing: as Gaussian primitives are repositioned or modified, their interpolated influence is immediately reflected in the rendered output. By combining the strengths of implicit and explicit representations, GENIE supports intuitive scene manipulation, dynamic interaction, and compatibility with physical simulation, bridging the gap between geometry-based editing and neural rendering. The code can be found under (https://github.com/MikolajZielinski/genie)
Authors: Anghong Du, Nay Aung, Theodoros N. Arvanitis, Stefan K. Piechnik, Joao A C Lima, Steffen E. Petersen, Le Zhang
Abstract: High-quality pixel-level annotations of medical images are essential for supervised segmentation tasks, but obtaining such annotations is costly and requires medical expertise. To address this challenge, we propose a novel coarse-to-fine segmentation framework that relies entirely on coarse-level annotations, encompassing both target and complementary drawings, despite their inherent noise. The framework works by introducing transition matrices in order to model the inaccurate and incomplete regions in the coarse annotations. By jointly training on multiple sets of coarse annotations, it progressively refines the network's outputs and infers the true segmentation distribution, achieving a robust approximation of precise labels through matrix-based modeling. To validate the flexibility and effectiveness of the proposed method, we demonstrate the results on two public cardiac imaging datasets, ACDC and MSCMRseg, and further evaluate its performance on the UK Biobank dataset. Experimental results indicate that our approach surpasses the state-of-the-art weakly supervised methods and closely matches the fully supervised approach.
Authors: Tianheng Zhu, Yiheng Feng
Abstract: As autonomous driving (AD) technology advances, increasing research has focused on leveraging cooperative perception (CP) data collected from multiple AVs to enhance traffic applications. Due to the impracticality of large-scale real-world AV deployments, simulation has become the primary approach in most studies. While game-engine-based simulators like CARLA generate high-fidelity raw sensor data (e.g., LiDAR point clouds) which can be used to produce realistic detection outputs, they face scalability challenges in multi-AV scenarios. In contrast, microscopic traffic simulators such as SUMO scale efficiently but lack perception modeling capabilities. To bridge this gap, we propose MIDAR, a LiDAR detection mimicking model that approximates realistic LiDAR detections using vehicle-level features readily available from microscopic traffic simulators. Specifically, MIDAR predicts true positives (TPs) and false negatives (FNs) from ideal LiDAR detection results based on the spatial layouts and dimensions of surrounding vehicles. A Refined Multi-hop Line-of-Sight (RM-LoS) graph is constructed to encode the occlusion relationships among vehicles, upon which MIDAR employs a GRU-enhanced APPNP architecture to propagate features from the ego AV and occluding vehicles to the prediction target. MIDAR achieves an AUC of 0.909 in approximating the detection results generated by CenterPoint, a mainstream 3D LiDAR detection model, on the nuScenes AD dataset. Two CP-based traffic applications further validate the necessity of such realistic detection modeling, particularly for tasks requiring accurate individual vehicle observations (e.g., position, speed, lane index). As demonstrated in the applications, MIDAR can be seamlessly integrated into traffic simulators and trajectory datasets and will be open-sourced upon publication.
Authors: J. Alex Hurt, Trevor M. Bajkowski, Grant J. Scott, Curt H. Davis
Abstract: In 2012, AlexNet established deep convolutional neural networks (DCNNs) as the state-of-the-art in CV, as these networks soon led in visual tasks for many domains, including remote sensing. With the publication of Visual Transformers, we are witnessing the second modern leap in computational vision, and as such, it is imperative to understand how various transformer-based neural networks perform on satellite imagery. While transformers have shown high levels of performance in natural language processing and CV applications, they have yet to be compared on a large scale to modern remote sensing data. In this paper, we explore the use of transformer-based neural networks for object detection in high-resolution electro-optical satellite imagery, demonstrating state-of-the-art performance on a variety of publicly available benchmark data sets. We compare eleven distinct bounding-box detection and localization algorithms in this study, of which seven were published since 2020, and all eleven since 2015. The performance of five transformer-based architectures is compared with six convolutional networks on three state-of-the-art opensource high-resolution remote sensing imagery datasets ranging in size and complexity. Following the training and evaluation of thirty-three deep neural models, we then discuss and analyze model performance across various feature extraction methodologies and detection algorithms.
Authors: Rongxin Jiang, Robert Long, Chenghao Gu, Mingrui Yan
Abstract: This paper introduces VisuCraft, a novel framework designed to significantly enhance the capabilities of Large Vision-Language Models (LVLMs) in complex visual-guided creative content generation. Existing LVLMs often exhibit limitations in maintaining high visual fidelity, genuine creativity, and precise adherence to nuanced user instructions when generating long-form texts. VisuCraft addresses these challenges by integrating a multimodal structured information extractor (E) and a dynamic prompt generation module (G). The extractor distills fine-grained visual attributes from input images into a rich, structured representation, which the dynamic prompt module then combines with user instructions to create highly optimized prompts for underlying LVLMs (e.g., LLaVA, InstructBLIP). Evaluated on the self-constructed ImageStoryGen-500K dataset using VisuGen Metrics (Visual Grounding, Creativity, and Instruction Adherence), VisuCraft consistently outperforms baseline LVLMs across tasks like story generation and poetry composition. Our results demonstrate remarkable improvements, particularly in creativity and instruction adherence, validating VisuCraft's effectiveness in producing imaginative, visually grounded, and user-aligned long-form creative text. This work unlocks new potential for LVLMs in sophisticated creative AI applications.
Authors: Mehrdad Moradi, Kamran Paynabar
Abstract: Recent advancements in diffusion models have demonstrated significant success in unsupervised anomaly segmentation. For anomaly segmentation, these models are first trained on normal data; then, an anomalous image is noised to an intermediate step, and the normal image is reconstructed through backward diffusion. Unlike traditional statistical methods, diffusion models do not rely on specific assumptions about the data or target anomalies, making them versatile for use across different domains. However, diffusion models typically assume access to normal data for training, limiting their applicability in realistic settings. In this paper, we propose novel robust denoising diffusion models for scenarios where only contaminated (i.e., a mix of normal and anomalous) unlabeled data is available. By casting maximum likelihood estimation of the data as a nonlinear regression problem, we reinterpret the denoising diffusion probabilistic model through a regression lens. Using robust regression, we derive a robust version of denoising diffusion probabilistic models. Our novel framework offers flexibility in constructing various robust diffusion models. Our experiments show that our approach outperforms current state of the art diffusion models, for unsupervised anomaly segmentation when only contaminated data is available. Our method outperforms existing diffusion-based approaches, achieving up to 8.08\% higher AUROC and 10.37\% higher AUPRC on MVTec datasets. The implementation code is available at: https://github.com/mehrdadmoradi124/RDDPM
Authors: Mahnoor Fatima Saad, Ziad Al-Halah
Abstract: How would the sound in a studio change with a carpeted floor and acoustic tiles on the walls? We introduce the task of material-controlled acoustic profile generation, where, given an indoor scene with specific audio-visual characteristics, the goal is to generate a target acoustic profile based on a user-defined material configuration at inference time. We address this task with a novel encoder-decoder approach that encodes the scene's key properties from an audio-visual observation and generates the target Room Impulse Response (RIR) conditioned on the material specifications provided by the user. Our model enables the generation of diverse RIRs based on various material configurations defined dynamically at inference time. To support this task, we create a new benchmark, the Acoustic Wonderland Dataset, designed for developing and evaluating material-aware RIR prediction methods under diverse and challenging settings. Our results demonstrate that the proposed model effectively encodes material information and generates high-fidelity RIRs, outperforming several baselines and state-of-the-art methods.
Authors: Vebj{\o}rn Haug K{\aa}sene, Pierre Lison
Abstract: Vision-and-Language Navigation (VLN) refers to the task of enabling autonomous robots to navigate unfamiliar environments by following natural language instructions. While recent Large Vision-Language Models (LVLMs) have shown promise in this task, most current VLM systems rely on models specifically designed and optimized for navigation, leaving the potential of off-the-shelf LVLMs underexplored. Furthermore, while older VLN approaches used low-level action spaces with egocentric views and atomic actions (such as "turn left" or "move forward"), newer models tend to favor panoramic action spaces with discrete navigable viewpoints. This paper investigates (1) whether off-the-shelf LVLMs (fine-tuned without architectural modifications or simulator-based training) can effectively support VLN tasks and (2) whether such models can support both low-level and panoramic action paradigms. To this end, we fine-tune the open-source model Qwen2.5-VL-3B-Instruct on the Room-to-Room (R2R) dataset and evaluate its empirical performance across both low-level and panoramic action spaces. The best resulting model achieves a 41% success rate on the R2R test set, demonstrating that while off-the-shelf LVLMs can learn to perform Vision-and-Language Navigation, they still lag behind models specifically designed for this task.
Authors: Minh-Hai Nguyen, Edouard Pauwels, Pierre Weiss
Abstract: The Maximum A Posteriori (MAP) estimation is a widely used framework in blind deconvolution to recover sharp images from blurred observations. The estimated image and blur filter are defined as the maximizer of the posterior distribution. However, when paired with sparsity-promoting image priors, MAP estimation has been shown to favors blurry solutions, limiting its effectiveness. In this paper, we revisit this result using diffusion-based priors, a class of models that capture realistic image distributions. Through an empirical examination of the prior's likelihood landscape, we uncover two key properties: first, blurry images tend to have higher likelihoods; second, the landscape contains numerous local minimizers that correspond to natural images. Building on these insights, we provide a theoretical analysis of the blind deblurring posterior. This reveals that the MAP estimator tends to produce sharp filters (close to the Dirac delta function) and blurry solutions. However local minimizers of the posterior, which can be obtained with gradient descent, correspond to realistic, natural images, effectively solving the blind deconvolution problem. Our findings suggest that overcoming MAP's limitations requires good local initialization to local minima in the posterior landscape. We validate our analysis with numerical experiments, demonstrating the practical implications of our insights for designing improved priors and optimization techniques.
Authors: Srikanth Muralidharan, Heitor R. Medeiros, Masih Aminbeidokhti, Eric Granger, Marco Pedersoli
Abstract: Many real-world applications require recognition models that are robust to different operational conditions and modalities, but at the same time run on small embedded devices, with limited hardware. While for normal size models, pre-training is known to be very beneficial in accuracy and robustness, for small models, that can be employed for embedded and edge devices, its effect is not clear. In this work, we investigate the effect of ImageNet pretraining on increasingly small backbone architectures (ultra-small models, with $<$1M parameters) with respect to robustness in downstream object detection tasks in the infrared visual modality. Using scaling laws derived from standard object recognition architectures, we construct two ultra-small backbone families and systematically study their performance. Our experiments on three different datasets reveal that while ImageNet pre-training is still useful, beyond a certain capacity threshold, it offers diminishing returns in terms of out-of-distribution detection robustness. Therefore, we advise practitioners to still use pre-training and, when possible avoid too small models as while they might work well for in-domain problems, they are brittle when working conditions are different.
Authors: Chenxu Zhang, Zenan Li, Hongyi Xu, You Xie, Xiaochen Zhao, Tianpei Gu, Guoxian Song, Xin Chen, Chao Liang, Jianwen Jiang, Linjie Luo
Abstract: We present X-Actor, a novel audio-driven portrait animation framework that generates lifelike, emotionally expressive talking head videos from a single reference image and an input audio clip. Unlike prior methods that emphasize lip synchronization and short-range visual fidelity in constrained speaking scenarios, X-Actor enables actor-quality, long-form portrait performance capturing nuanced, dynamically evolving emotions that flow coherently with the rhythm and content of speech. Central to our approach is a two-stage decoupled generation pipeline: an audio-conditioned autoregressive diffusion model that predicts expressive yet identity-agnostic facial motion latent tokens within a long temporal context window, followed by a diffusion-based video synthesis module that translates these motions into high-fidelity video animations. By operating in a compact facial motion latent space decoupled from visual and identity cues, our autoregressive diffusion model effectively captures long-range correlations between audio and facial dynamics through a diffusion-forcing training paradigm, enabling infinite-length emotionally-rich motion prediction without error accumulation. Extensive experiments demonstrate that X-Actor produces compelling, cinematic-style performances that go beyond standard talking head animations and achieves state-of-the-art results in long-range, audio-driven emotional portrait acting.
Authors: Dawei Zhang, Xiaojie Guo
Abstract: Despite notable advances in image denoising, existing models often struggle to generalize beyond in-distribution noise patterns, particularly when confronted with out-of-distribution (OOD) conditions characterized by spatially variant noise. This generalization gap remains a fundamental yet underexplored challenge. In this work, we investigate \emph{scale equivariance} as a core inductive bias for improving OOD robustness. We argue that incorporating scale-equivariant structures enables models to better adapt from training on spatially uniform noise to inference on spatially non-uniform degradations. Building on this insight, we propose a robust blind denoising framework equipped with two key components: a Heterogeneous Normalization Module (HNM) and an Interactive Gating Module (IGM). HNM stabilizes feature distributions and dynamically corrects features under varying noise intensities, while IGM facilitates effective information modulation via gated interactions between signal and feature paths. Extensive evaluations demonstrate that our model consistently outperforms state-of-the-art methods on both synthetic and real-world benchmarks, especially under spatially heterogeneous noise. Code will be made publicly available.
Authors: Alakh Desai, Nuno Vasconcelos
Abstract: Diffusion models (DMs) have demonstrated an unparalleled ability to create diverse and high-fidelity images from text prompts. However, they are also well-known to vary substantially regarding both prompt adherence and quality. Negative prompting was introduced to improve prompt compliance by specifying what an image must not contain. Previous works have shown the existence of an ideal negative prompt that can maximize the odds of the positive prompt. In this work, we explore relations between negative prompting and classifier-free guidance (CFG) to develop a sampling procedure, {\it Adaptive Negative Sampling Without External Resources} (ANSWER), that accounts for both positive and negative conditions from a single prompt. This leverages the internal understanding of negation by the diffusion model to increase the odds of generating images faithful to the prompt. ANSWER is a training-free technique, applicable to any model that supports CFG, and allows for negative grounding of image concepts without an explicit negative prompts, which are lossy and incomplete. Experiments show that adding ANSWER to existing DMs outperforms the baselines on multiple benchmarks and is preferred by humans 2x more over the other methods.
Authors: Yusaku Takama, Ning Ding, Tatsuya Yokota, Toru Tamaki
Abstract: Existing architectures of multi-domain learning have two types of adapters: shared LoRA for all domains and domain-specific LoRA for each particular domain. However, it remains unclear whether this structure effectively captures domain-specific information. In this paper, we propose a method that ensures that shared and domain-specific LoRAs exist in different subspaces; specifically, the column and left null subspaces of the pre-trained weights. We apply the proposed method to action recognition with three datasets (UCF101, Kinetics400, and HMDB51) and demonstrate its effectiveness in some cases along with the analysis of the dimensions of LoRA weights.
Authors: Takuya Sugimoto, Ning Ding, Toru Tamaki
Abstract: Modern action recognition models suffer from static bias, leading to reduced generalization performance. In this paper, we propose MoExDA, a lightweight domain adaptation between RGB and edge information using edge frames in addition to RGB frames to counter the static bias issue. Experiments demonstrate that the proposed method effectively suppresses static bias with a lower computational cost, allowing for more robust action recognition than previous approaches.
Authors: Zachary Yahn, Selim Furkan Tekin, Fatih Ilhan, Sihao Hu, Tiansheng Huang, Yichang Xu, Margaret Loper, Ling Liu
Abstract: Adversarial perturbations are useful tools for exposing vulnerabilities in neural networks. Existing adversarial perturbation methods for object detection are either limited to attacking CNN-based detectors or weak against transformer-based detectors. This paper presents an Attention-Focused Offensive Gradient (AFOG) attack against object detection transformers. By design, AFOG is neural-architecture agnostic and effective for attacking both large transformer-based object detectors and conventional CNN-based detectors with a unified adversarial attention framework. This paper makes three original contributions. First, AFOG utilizes a learnable attention mechanism that focuses perturbations on vulnerable image regions in multi-box detection tasks, increasing performance over non-attention baselines by up to 30.6%. Second, AFOG's attack loss is formulated by integrating two types of feature loss through learnable attention updates with iterative injection of adversarial perturbations. Finally, AFOG is an efficient and stealthy adversarial perturbation method. It probes the weak spots of detection transformers by adding strategically generated and visually imperceptible perturbations which can cause well-trained object detection models to fail. Extensive experiments conducted with twelve large detection transformers on COCO demonstrate the efficacy of AFOG. Our empirical results also show that AFOG outperforms existing attacks on transformer-based and CNN-based object detectors by up to 83% with superior speed and imperceptibility. Code is available at https://github.com/zacharyyahn/AFOG.
Authors: Fan Yang, Yihao Huang, Jiayi Zhu, Ling Shi, Geguang Pu, Jin Song Dong, Kailong Wang
Abstract: Diffusion-based text-to-image (T2I) models enable high-quality image generation but also pose significant risks of misuse, particularly in producing not-safe-for-work (NSFW) content. While prior detection methods have focused on filtering prompts before generation or moderating images afterward, the in-generation phase of diffusion models remains largely unexplored for NSFW detection. In this paper, we introduce In-Generation Detection (IGD), a simple yet effective approach that leverages the predicted noise during the diffusion process as an internal signal to identify NSFW content. This approach is motivated by preliminary findings suggesting that the predicted noise may capture semantic cues that differentiate NSFW from benign prompts, even when the prompts are adversarially crafted. Experiments conducted on seven NSFW categories show that IGD achieves an average detection accuracy of 91.32% over naive and adversarial NSFW prompts, outperforming seven baseline methods.
Authors: Xinhui Li, Xiaojie Guo
Abstract: Domain Generalized Semantic Segmentation (DGSS) aims to improve the generalization ability of models across unseen domains without access to target data during training. Recent advances in DGSS have increasingly exploited vision foundation models (VFMs) via parameter-efficient fine-tuning strategies. However, most existing approaches concentrate on global feature fine-tuning, while overlooking hierarchical adaptation across feature levels, which is crucial for precise dense prediction. In this paper, we propose Multi-Granularity Feature Calibration (MGFC), a novel framework that performs coarse-to-fine alignment of VFM features to enhance robustness under domain shifts. Specifically, MGFC first calibrates coarse-grained features to capture global contextual semantics and scene-level structure. Then, it refines medium-grained features by promoting category-level feature discriminability. Finally, fine-grained features are calibrated through high-frequency spatial detail enhancement. By performing hierarchical and granularity-aware calibration, MGFC effectively transfers the generalization strengths of VFMs to the domain-specific task of DGSS. Extensive experiments on benchmark datasets demonstrate that our method outperforms state-of-the-art DGSS approaches, highlighting the effectiveness of multi-granularity adaptation for the semantic segmentation task of domain generalization.
Authors: Xuyi Yang, Wenhao Zhang, Hongbo Jin, Lin Liu, Hongbo Xu, Yongwei Nie, Fei Yu, Fei Ma
Abstract: Current Multimodal Large Language Models (MLLMs) often perform poorly in long video understanding, primarily due to resource limitations that prevent them from processing all video frames and their associated information. Efficiently extracting relevant information becomes a challenging task. Existing frameworks and evaluation tasks focus on identifying specific frames containing core objects from a large number of irrelevant frames, which does not align with the practical needs of real-world applications. To address this issue, we propose a new scenario under the video question-answering task, SceneQA, which emphasizes scene-based detail perception and reasoning abilities. And we develop the LVSQA dataset to support the SceneQA task, which is built upon carefully selected videos from LVBench and contains a new collection of question-answer pairs to promote a more fair evaluation of MLLMs' scene perception abilities in long videos. Inspired by human cognition, we introduce a novel method called SLFG. The core idea of SLFG is to combine individual frames into semantically coherent scene frames. By leveraging scene localization methods and dynamic frame reassembly mechanisms, SLFG significantly enhances the understanding capabilities of existing MLLMs in long videos. SLFG requires no modification to the original model architecture and boasts excellent plug-and-play usability. Experimental results show that this method performs exceptionally well in several long video benchmark tests. Code and dataset will be released at http://www.slfg.pkuzwh.cn.
Authors: Liheng Zhang, Weihao Yu, Zubo Lu, Haozhi Gu, Jin Huang
Abstract: Recent advancements in 3D Gaussian Splatting have enhanced efficient and high-quality novel view synthesis. However, representing scenes requires a large number of Gaussian points, leading to high storage demands and limiting practical deployment. The latest methods facilitate the compression of Gaussian models but struggle to identify truly insignificant Gaussian points in the scene, leading to a decline in subsequent Gaussian pruning, compression quality, and rendering performance. To address this issue, we propose SA-3DGS, a method that significantly reduces storage costs while maintaining rendering quality. SA-3DGS learns an importance score to automatically identify the least significant Gaussians in scene reconstruction, thereby enabling effective pruning and redundancy reduction. Next, the importance-aware clustering module compresses Gaussians attributes more accurately into the codebook, improving the codebook's expressive capability while reducing model size. Finally, the codebook repair module leverages contextual scene information to repair the codebook, thereby recovering the original Gaussian point attributes and mitigating the degradation in rendering quality caused by information loss. Experimental results on several benchmark datasets show that our method achieves up to 66x compression while maintaining or even improving rendering quality. The proposed Gaussian pruning approach is not only adaptable to but also improves other pruning-based methods (e.g., LightGaussian), showcasing excellent performance and strong generalization ability.
Authors: Qi Xie (University of Science and Technology of China), Yongjia Ma (Li Auto), Donglin Di (Li Auto), Xuehao Gao (Northwestern Polytechnical University), Xun Yang (University of Science and Technology of China)
Abstract: Achieving ID-preserving text-to-video (T2V) generation remains challenging despite recent advances in diffusion-based models. Existing approaches often fail to capture fine-grained facial dynamics or maintain temporal identity coherence. To address these limitations, we propose MoCA, a novel Video Diffusion Model built on a Diffusion Transformer (DiT) backbone, incorporating a Mixture of Cross-Attention mechanism inspired by the Mixture-of-Experts paradigm. Our framework improves inter-frame identity consistency by embedding MoCA layers into each DiT block, where Hierarchical Temporal Pooling captures identity features over varying timescales, and Temporal-Aware Cross-Attention Experts dynamically model spatiotemporal relationships. We further incorporate a Latent Video Perceptual Loss to enhance identity coherence and fine-grained details across video frames. To train this model, we collect CelebIPVid, a dataset of 10,000 high-resolution videos from 1,000 diverse individuals, promoting cross-ethnicity generalization. Extensive experiments on CelebIPVid show that MoCA outperforms existing T2V methods by over 5% across Face similarity.
Authors: Yiran Meng, Junhong Ye, Wei Zhou, Guanghui Yue, Xudong Mao, Ruomei Wang, Baoquan Zhao
Abstract: Cross-video question answering presents significant challenges beyond traditional single-video understanding, particularly in establishing meaningful connections across video streams and managing the complexity of multi-source information retrieval. We introduce VideoForest, a novel framework that addresses these challenges through person-anchored hierarchical reasoning. Our approach leverages person-level features as natural bridge points between videos, enabling effective cross-video understanding without requiring end-to-end training. VideoForest integrates three key innovations: 1) a human-anchored feature extraction mechanism that employs ReID and tracking algorithms to establish robust spatiotemporal relationships across multiple video sources; 2) a multi-granularity spanning tree structure that hierarchically organizes visual content around person-level trajectories; and 3) a multi-agent reasoning framework that efficiently traverses this hierarchical structure to answer complex cross-video queries. To evaluate our approach, we develop CrossVideoQA, a comprehensive benchmark dataset specifically designed for person-centric cross-video analysis. Experimental results demonstrate VideoForest's superior performance in cross-video reasoning tasks, achieving 71.93% accuracy in person recognition, 83.75% in behavior analysis, and 51.67% in summarization and reasoning, significantly outperforming existing methods. Our work establishes a new paradigm for cross-video understanding by unifying multiple video streams through person-level features, enabling sophisticated reasoning across distributed visual information while maintaining computational efficiency.
Authors: Zeyu Zhu, Weijia Wu, Mike Zheng Shou
Abstract: Existing studies on talking video generation have predominantly focused on single-person monologues or isolated facial animations, limiting their applicability to realistic multi-human interactions. To bridge this gap, we introduce MIT, a large-scale dataset specifically designed for multi-human talking video generation. To this end, we develop an automatic pipeline that collects and annotates multi-person conversational videos. The resulting dataset comprises 12 hours of high-resolution footage, each featuring two to four speakers, with fine-grained annotations of body poses and speech interactions. It captures natural conversational dynamics in multi-speaker scenario, offering a rich resource for studying interactive visual behaviors. To demonstrate the potential of MIT, we furthur propose CovOG, a baseline model for this novel task. It integrates a Multi-Human Pose Encoder (MPE) to handle varying numbers of speakers by aggregating individual pose embeddings, and an Interactive Audio Driver (IAD) to modulate head dynamics based on speaker-specific audio features. Together, these components showcase the feasibility and challenges of generating realistic multi-human talking videos, establishing MIT as a valuable benchmark for future research. The code is avalibale at: https://github.com/showlab/Multi-human-Talking-Video-Dataset.
URLs: https://github.com/showlab/Multi-human-Talking-Video-Dataset.
Authors: Hyebin Cho, Jaehyup Lee
Abstract: Face filters have become a key element of short-form video content, enabling a wide array of visual effects such as stylization and face swapping. However, their performance often degrades in the presence of occlusions, where objects like hands, hair, or accessories obscure the face. To address this limitation, we introduce the novel task of face matting, which estimates fine-grained alpha mattes to separate occluding elements from facial regions. We further present FaceMat, a trimap-free, uncertainty-aware framework that predicts high-quality alpha mattes under complex occlusions. Our approach leverages a two-stage training pipeline: a teacher model is trained to jointly estimate alpha mattes and per-pixel uncertainty using a negative log-likelihood (NLL) loss, and this uncertainty is then used to guide the student model through spatially adaptive knowledge distillation. This formulation enables the student to focus on ambiguous or occluded regions, improving generalization and preserving semantic consistency. Unlike previous approaches that rely on trimaps or segmentation masks, our framework requires no auxiliary inputs making it well-suited for real-time applications. In addition, we reformulate the matting objective by explicitly treating skin as foreground and occlusions as background, enabling clearer compositing strategies. To support this task, we newly constructed CelebAMat, a large-scale synthetic dataset specifically designed for occlusion-aware face matting. Extensive experiments show that FaceMat outperforms state-of-the-art methods across multiple benchmarks, enhancing the visual quality and robustness of face filters in real-world, unconstrained video scenarios. The source code and CelebAMat dataset are available at https://github.com/hyebin-c/FaceMat.git
Authors: Lekang Wen, Jing Xiao, Liang Liao, Jiajun Chen, Mi Wang
Abstract: Modality-agnostic Semantic Segmentation (MaSS) aims to achieve robust scene understanding across arbitrary combinations of input modality. Existing methods typically rely on explicit feature alignment to achieve modal homogenization, which dilutes the distinctive strengths of each modality and destroys their inherent complementarity. To achieve cooperative harmonization rather than homogenization, we propose CHARM, a novel complementary learning framework designed to implicitly align content while preserving modality-specific advantages through two components: (1) Mutual Perception Unit (MPU), enabling implicit alignment through window-based cross-modal interaction, where modalities serve as both queries and contexts for each other to discover modality-interactive correspondences; (2) A dual-path optimization strategy that decouples training into Collaborative Learning Strategy (CoL) for complementary fusion learning and Individual Enhancement Strategy (InE) for protected modality-specific optimization. Experiments across multiple datasets and backbones indicate that CHARM consistently outperform the baselines, with significant increment on the fragile modalities. This work shifts the focus from model homogenization to harmonization, enabling cross-modal complementarity for true harmony in diversity.
Authors: Trinh Quoc Nguyen, Oky Dicky Ardiansyah Prima, Katsuyoshi Hotta
Abstract: This study introduces a novel framework, "Comprehensive Optimization and Refinement through Ensemble Fusion in Domain Adaptation for Person Re-identification (CORE-ReID)", to address an Unsupervised Domain Adaptation (UDA) for Person Re-identification (ReID). The framework utilizes CycleGAN to generate diverse data that harmonizes differences in image characteristics from different camera sources in the pre-training stage. In the fine-tuning stage, based on a pair of teacher-student networks, the framework integrates multi-view features for multi-level clustering to derive diverse pseudo labels. A learnable Ensemble Fusion component that focuses on fine-grained local information within global features is introduced to enhance learning comprehensiveness and avoid ambiguity associated with multiple pseudo-labels. Experimental results on three common UDAs in Person ReID demonstrate significant performance gains over state-of-the-art approaches. Additional enhancements, such as Efficient Channel Attention Block and Bidirectional Mean Feature Normalization mitigate deviation effects and adaptive fusion of global and local features using the ResNet-based model, further strengthening the framework. The proposed framework ensures clarity in fusion features, avoids ambiguity, and achieves high ac-curacy in terms of Mean Average Precision, Top-1, Top-5, and Top-10, positioning it as an advanced and effective solution for the UDA in Person ReID. Our codes and models are available at https://github.com/TrinhQuocNguyen/CORE-ReID.
Authors: Bo Zhang, Yifan Zhang, Shuo Yan, Yu Bai, Zheng Zhang, Wu Liu, Xiuzhuang Zhou, Wendong Wang
Abstract: In light of the spatial domain's limited capacity for modeling global context in 3D medical image segmentation, emerging approaches have begun to incorporate frequency domain representations. However, straightforward feature extraction strategies often overlook the unique properties of frequency domain information, such as conjugate symmetry. They also fail to account for the fundamental differences in data distribution between the spatial and frequency domains, which can ultimately dilute or obscure the complementary strengths that frequency-based representations offer. In this paper, we propose SSFMamba, a Mamba based Symmetry-driven Spatial-Frequency feature fusion network for 3D medical image segmentation. SSFMamba employs a complementary dual-branch architecture that extracts features from both the spatial and frequency domains, and leverages a Mamba block to fuse these heterogeneous features to preserve global context while reinforcing local details. In the frequency domain branch, we harness Mamba's exceptional capability to extract global contextual information in conjunction with the synergistic effect of frequency domain features to further enhance global modeling. Moreover, we design a 3D multi-directional scanning mechanism to strengthen the fusion of local and global cues. Extensive experiments on the BraTS2020 and BraTS2023 datasets demonstrate that our approach consistently outperforms state-of-the-art methods across various evaluation metrics.
Authors: Anran Wu, Long Peng, Xin Di, Xueyuan Dai, Chen Wu, Yang Wang, Xueyang Fu, Yang Cao, Zheng-Jun Zha
Abstract: Feedforward 3D Gaussian Splatting (3DGS) overcomes the limitations of optimization-based 3DGS by enabling fast and high-quality reconstruction without the need for per-scene optimization. However, existing feedforward approaches typically assume that input multi-view images are clean and high-quality. In real-world scenarios, images are often captured under challenging conditions such as noise, low light, or rain, resulting in inaccurate geometry and degraded 3D reconstruction. To address these challenges, we propose a general and efficient multi-view feature enhancement module, RobustGS, which substantially improves the robustness of feedforward 3DGS methods under various adverse imaging conditions, enabling high-quality 3D reconstruction. The RobustGS module can be seamlessly integrated into existing pretrained pipelines in a plug-and-play manner to enhance reconstruction robustness. Specifically, we introduce a novel component, Generalized Degradation Learner, designed to extract generic representations and distributions of multiple degradations from multi-view inputs, thereby enhancing degradation-awareness and improving the overall quality of 3D reconstruction. In addition, we propose a novel semantic-aware state-space model. It first leverages the extracted degradation representations to enhance corrupted inputs in the feature space. Then, it employs a semantic-aware strategy to aggregate semantically similar information across different views, enabling the extraction of fine-grained cross-view correspondences and further improving the quality of 3D representations. Extensive experiments demonstrate that our approach, when integrated into existing methods in a plug-and-play manner, consistently achieves state-of-the-art reconstruction quality across various types of degradations.
Authors: Zaiying Zhao, Toshihiko Yamasaki
Abstract: The rapid expansion of applications using Large Vision-Language Models (LVLMs), such as GPT-4o, has raised significant concerns about their fairness. While existing studies primarily focus on demographic attributes such as race and gender, fairness across a broader range of attributes remains largely unexplored. In this study, we construct an open-set knowledge base of bias attributes leveraging Large Language Models (LLMs) and evaluate the fairness of LVLMs across finer-grained attributes. Our experimental results reveal that LVLMs exhibit biased outputs across a diverse set of attributes and further demonstrate that cultural, environmental, and behavioral factors have a more pronounced impact on LVLM decision-making than traditional demographic attributes.
Authors: Bo Zhang, Xu Xinan, Shuo Yan, Yu Bai, Zheng Zhang, Wufan Wang, Wendong Wang
Abstract: Recent pseudo-bag augmentation methods for Multiple Instance Learning (MIL)-based Whole Slide Image (WSI) classification sample instances from a limited number of bags, resulting in constrained diversity. To address this issue, we propose Contrastive Cross-Bag Augmentation ($C^2Aug$) to sample instances from all bags with the same class to increase the diversity of pseudo-bags. However, introducing new instances into the pseudo-bag increases the number of critical instances (e.g., tumor instances). This increase results in a reduced occurrence of pseudo-bags containing few critical instances, thereby limiting model performance, particularly on test slides with small tumor areas. To address this, we introduce a bag-level and group-level contrastive learning framework to enhance the discrimination of features with distinct semantic meanings, thereby improving model performance. Experimental results demonstrate that $C^2Aug$ consistently outperforms state-of-the-art approaches across multiple evaluation metrics.
Authors: Jiantao Tan, Peixian Ma, Kanghao Chen, Zhiming Dai, Ruixuan Wang
Abstract: Continual learning is essential for medical image classification systems to adapt to dynamically evolving clinical environments. The integration of multimodal information can significantly enhance continual learning of image classes. However, while existing approaches do utilize textual modality information, they solely rely on simplistic templates with a class name, thereby neglecting richer semantic information. To address these limitations, we propose a novel framework that harnesses visual concepts generated by large language models (LLMs) as discriminative semantic guidance. Our method dynamically constructs a visual concept pool with a similarity-based filtering mechanism to prevent redundancy. Then, to integrate the concepts into the continual learning process, we employ a cross-modal image-concept attention module, coupled with an attention loss. Through attention, the module can leverage the semantic knowledge from relevant visual concepts and produce class-representative fused features for classification. Experiments on medical and natural image datasets show our method achieves state-of-the-art performance, demonstrating the effectiveness and superiority of our method. We will release the code publicly.
Authors: Yogesh Kulkarni, Pooyan Fazli
Abstract: Multimodal reasoning over long-horizon video is challenging due to the need for precise spatiotemporal fusion and alignment across modalities. While recent methods such as Group Relative Policy Optimization (GRPO) have shown promise in this domain, they suffer from three key limitations: (1) data inefficiency from their on-policy design, (2) a vanishing advantage problem, where identical or near-identical rewards within a group eliminate the learning signal by producing zero-valued advantages, and (3) uniform credit assignment that fails to emphasize critical reasoning steps. We introduce AVATAR (Audio-Video Agent for Alignment and Reasoning), a framework that addresses these limitations through two core components: (1) an off-policy training architecture that improves sample efficiency and resolves vanishing advantages by reusing past experiences with greater reward diversity, and (2) Temporal Advantage Shaping (TAS), a novel credit assignment strategy that upweights key reasoning phases during learning. AVATAR achieves strong performance across various benchmarks, outperforming the Qwen2.5-Omni baseline by +5.4on MMVU, +4.9 on OmniBench, and +4.5 on Video-Holmes, while demonstrating over 35% higher sample efficiency.
Authors: Tianjiao Jiang, Zhen Zhang, Yuhang Liu, Javen Qinfeng Shi
Abstract: Few-shot learning (FSL) often requires effective adaptation of models using limited labeled data. However, most existing FSL methods rely on entangled representations, requiring the model to implicitly recover the unmixing process to obtain disentangled representations using only limited supervision, which hinders effective adaptation. Recent theoretical studies show that multimodal contrastive learning methods, such as CLIP, can disentangle latent representations up to linear transformations. In light of this, we propose the Causal CLIP Adapter (CCA), a novel framework that explicitly disentangles visual features extracted from CLIP using unsupervised Independent Component Analysis (ICA). This removes the need to learn the unmixing process from the labeled data, thereby reducing the number of trainable parameters and mitigating overfitting. Taking a step further, while ICA can obtain visual disentangled representations, it may also disrupt CLIP's intra- and inter-modal alignment. To counteract this, CCA further leverages CLIP's inherent cross-modal alignment by enhancing it in two ways: unidirectionally, through fine-tuning a CLIP-based text classifier, and bidirectionally, via a cross-attention mechanism that enriches visual and textual representations through mutual interaction. Both unimodal and cross-modal classification outputs can be effectively combined linearly to improve classification accuracy. Extensive experiments on 11 benchmark datasets demonstrate that our method consistently outperforms state-of-the-art approaches in terms of few-shot performance and robustness to distributional shifts, while maintaining computational efficiency. Code will be available at https://github.com/tianjiao-j/CCA.
Authors: Heng Jia, Linchao Zhu, Na Zhao
Abstract: Despite recent advances in feed-forward 3D Gaussian Splatting, generalizable 3D reconstruction remains challenging, particularly in multi-view correspondence modeling. Existing approaches face a fundamental trade-off: explicit methods achieve geometric precision but struggle with ambiguous regions, while implicit methods provide robustness but suffer from slow convergence. We present H3R, a hybrid framework that addresses this limitation by integrating volumetric latent fusion with attention-based feature aggregation. Our framework consists of two complementary components: an efficient latent volume that enforces geometric consistency through epipolar constraints, and a camera-aware Transformer that leverages Pl\"ucker coordinates for adaptive correspondence refinement. By integrating both paradigms, our approach enhances generalization while converging 2$\times$ faster than existing methods. Furthermore, we show that spatial-aligned foundation models (e.g., SD-VAE) substantially outperform semantic-aligned models (e.g., DINOv2), resolving the mismatch between semantic representations and spatial reconstruction requirements. Our method supports variable-number and high-resolution input views while demonstrating robust cross-dataset generalization. Extensive experiments show that our method achieves state-of-the-art performance across multiple benchmarks, with significant PSNR improvements of 0.59 dB, 1.06 dB, and 0.22 dB on the RealEstate10K, ACID, and DTU datasets, respectively. Code is available at https://github.com/JiaHeng-DLUT/H3R.
Authors: Sai Ma, Zhuang Li, John A Taylor
Abstract: Vision language models (VLMs) that enable natural language interaction with satellite imagery can democratize Earth observation by accelerating expert workflows, making data accessible to non-specialists, and enabling planet-scale automation. However, existing datasets focus mainly on short-term, high-resolution imagery from a limited number of satellites, overlooking low-resolution, multi-satellite, long-term archives, such as Landsat, that are essential for affordable and bias-robust global monitoring. We address this gap with Landsat30-AU, a large-scale vision-language dataset built from 30-meter resolution imagery collected by four Landsat satellites (5, 7, 8, and 9) over Australia, spanning more than 36 years. The dataset includes two components: Landsat30-AU-Cap, containing $196,262$ image-caption pairs, and Landsat30-AU-VQA, comprising 17,725 human-verified visual question answering (VQA) samples across eight remote sensing domains. Both datasets are curated through a bootstrapped pipeline that leverages generic VLMs with iterative refinement and human verification to ensure quality. Our evaluation of eight VLMs on our benchmark reveals that off-the-shelf models struggle to understand satellite imagery. The open-source remote-sensing VLM EarthDial achieves only 0.07 SPIDEr in captioning and a VQA accuracy of 0.48, highlighting the limitations of current approaches. Encouragingly, lightweight fine-tuning of Qwen2.5-VL-7B on Landsat30-AU improves captioning performance from 0.11 to 0.31 SPIDEr and boosts VQA accuracy from 0.74 to 0.87. Code and data are available at https://github.com/papersubmit1/landsat30-au.
Authors: Arion Zimmermann, Soon-Jo Chung, Fred Hadaegh
Abstract: The accurate state estimation of unknown bodies in space is a critical challenge with applications ranging from the tracking of space debris to the shape estimation of small bodies. A necessary enabler to this capability is to find and track features on a continuous stream of images. Existing methods, such as SIFT, ORB and AKAZE, achieve real-time but inaccurate pose estimates, whereas modern deep learning methods yield higher quality features at the cost of more demanding computational resources which might not be available on space-qualified hardware. Additionally, both classical and data-driven methods are not robust to the highly opaque self-cast shadows on the object of interest. We show that, as the target body rotates, these shadows may lead to large biases in the resulting pose estimates. For these objects, a bias in the real-time pose estimation algorithm may mislead the spacecraft's state estimator and cause a mission failure, especially if the body undergoes a chaotic tumbling motion. We present COFFEE, the Celestial Occlusion Fast FEature Extractor, a real-time pose estimation framework for asteroids designed to leverage prior information on the sun phase angle given by sun-tracking sensors commonly available onboard spacecraft. By associating salient contours to their projected shadows, a sparse set of features are detected, invariant to the motion of the shadows. A Sparse Neural Network followed by an attention-based Graph Neural Network feature matching model are then jointly trained to provide a set of correspondences between successive frames. The resulting pose estimation pipeline is found to be bias-free, more accurate than classical pose estimation pipelines and an order of magnitude faster than other state-of-the-art deep learning pipelines on synthetic data as well as on renderings of the tumbling asteroid Apophis.
Authors: Haozhou Zhai, Yanzhe Gao, Tianjiang Hu
Abstract: Fire scene datasets are crucial for training robust computer vision models, particularly in tasks such as fire early warning and emergency rescue operations. However, among the currently available fire-related data, there is a significant shortage of annotated data specifically targeting building units.To tackle this issue, we introduce an annotated dataset of building units captured by drones, which incorporates multiple enhancement techniques. We construct backgrounds using real multi-story scenes, combine motion blur and brightness adjustment to enhance the authenticity of the captured images, simulate drone shooting conditions under various circumstances, and employ large models to generate fire effects at different locations.The synthetic dataset generated by this method encompasses a wide range of building scenarios, with a total of 1,978 images. This dataset can effectively improve the generalization ability of fire unit detection, providing multi-scenario and scalable data while reducing the risks and costs associated with collecting real fire data. The dataset is available at https://github.com/boilermakerr/FireUnitData.
Authors: Chengyu Bai, Jintao Chen, Xiang Bai, Yilong Chen, Qi She, Ming Lu, Shanghang Zhang
Abstract: In recent years, unified vision-language models (VLMs) have rapidly advanced, effectively tackling both visual understanding and generation tasks within a single design. While many unified VLMs have explored various design choices, the recent hypothesis from OpenAI's GPT-4o suggests a promising generation pipeline: Understanding VLM->Visual Feature->Projector->Diffusion Model->Image. The understanding VLM is frozen, and only the generation-related modules are trained. This pipeline maintains the strong capability of understanding VLM while enabling the image generation ability of the unified VLM. Although this pipeline has shown very promising potential for the future development of unified VLM, how to easily enable image editing capability is still unexplored. In this paper, we introduce a novel training-free framework named UniEdit-I to enable the unified VLM with image editing capability via three iterative steps: understanding, editing, and verifying. 1. The understanding step analyzes the source image to create a source prompt through structured semantic analysis and makes minimal word replacements to form the target prompt based on the editing instruction. 2. The editing step introduces a time-adaptive offset, allowing for coherent editing from coarse to fine throughout the denoising process. 3. The verification step checks the alignment between the target prompt and the intermediate edited image, provides automatic consistency scores and corrective feedback, and determines whether to stop early or continue the editing loop. This understanding, editing, and verifying loop iterates until convergence, delivering high-fidelity editing in a training-free manner. We implemented our method based on the latest BLIP3-o and achieved state-of-the-art (SOTA) performance on the GEdit-Bench benchmark.
Authors: Yanshu Wang, Xichen Xu, Xiaoning Lei, Guoyang Xie
Abstract: Synthesizing realistic and spatially precise anomalies is essential for enhancing the robustness of industrial anomaly detection systems. While recent diffusion-based methods have demonstrated strong capabilities in modeling complex defect patterns, they often struggle with spatial controllability and fail to maintain fine-grained regional fidelity. To overcome these limitations, we propose SARD (Segmentation-Aware anomaly synthesis via Region-constrained Diffusion with discriminative mask Guidance), a novel diffusion-based framework specifically designed for anomaly generation. Our approach introduces a Region-Constrained Diffusion (RCD) process that preserves the background by freezing it and selectively updating only the foreground anomaly regions during the reverse denoising phase, thereby effectively reducing background artifacts. Additionally, we incorporate a Discriminative Mask Guidance (DMG) module into the discriminator, enabling joint evaluation of both global realism and local anomaly fidelity, guided by pixel-level masks. Extensive experiments on the MVTec-AD and BTAD datasets show that SARD surpasses existing methods in segmentation accuracy and visual quality, setting a new state-of-the-art for pixel-level anomaly synthesis.
Authors: Liangyang Ouyang, Jiafeng Mao
Abstract: Text-driven image editing enables users to flexibly modify visual content through natural language instructions, and is widely applied to tasks such as semantic object replacement, insertion, and removal. While recent inversion-based editing methods using rectified flow models have achieved promising results in image quality, we identify a structural limitation in their editing behavior: the semantic bias toward the source concept encoded in the inverted noise tends to suppress attention to the target concept. This issue becomes particularly critical when the source and target semantics are dissimilar, where the attention mechanism inherently leads to editing failure or unintended modifications in non-target regions. In this paper, we systematically analyze and validate this structural flaw, and introduce LORE, a training-free and efficient image editing method. LORE directly optimizes the inverted noise, addressing the core limitations in generalization and controllability of existing approaches, enabling stable, controllable, and general-purpose concept replacement, without requiring architectural modification or model fine-tuning. We conduct comprehensive evaluations on three challenging benchmarks: PIEBench, SmartEdit, and GapEdit. Experimental results show that LORE significantly outperforms strong baselines in terms of semantic alignment, image quality, and background fidelity, demonstrating the effectiveness and scalability of latent-space optimization for general-purpose image editing.
Authors: Junyoung Lim, Jaewoo Ahn, Gunhee Kim
Abstract: Generating accurate, informative, and hallucination-free captions for charts remains challenging for vision language models, primarily due to the lack of large-scale, high-quality datasets of real-world charts. However, existing real-world chart datasets suffer from the inclusion of extraneous information that cannot be inferred from the chart and failure to sufficiently capture structural elements and key insights. Therefore, we introduce ChartCap, a large-scale dataset of 565K real-world chart images paired with type-specific, dense captions that exclude extraneous information and highlight both structural elements and key insights in detail. To build ChartCap, we design a four-stage pipeline that generates captions using only the discernible data from the chart and employ a cycle consistency-based human verification, which accelerates quality control without sacrificing accuracy. Additionally, we propose a novel metric, the Visual Consistency Score, which evaluates caption quality by measuring the similarity between the chart regenerated from a caption and the original chart, independent of reference captions. Extensive experiments confirms that models fine-tuned on ChartCap consistently generate more accurate and informative captions with reduced hallucinations, surpassing both open-source and proprietary models and even human-annotated captions.
Authors: Zhaoxu Li, Chenqi Kong, Yi Yu, Qiangqiang Wu, Xinghao Jiang, Ngai-Man Cheung, Bihan Wen, Alex Kot, Xudong Jiang
Abstract: Large Vision-Language Models (LVLMs) recently achieve significant breakthroughs in understanding complex visual-textual contexts. However, hallucination issues still limit their real-world applicability. Although previous mitigation methods effectively reduce hallucinations in photographic images, they largely overlook the potential risks posed by stylized images, which play crucial roles in critical scenarios such as game scene understanding, art education, and medical analysis. In this work, we first construct a dataset comprising photographic images and their corresponding stylized versions with carefully annotated caption labels. We then conduct head-to-head comparisons on both discriminative and generative tasks by benchmarking 13 advanced LVLMs on the collected datasets. Our findings reveal that stylized images tend to induce significantly more hallucinations than their photographic counterparts. To address this issue, we propose Style-Aware Visual Early Revision SAVER, a novel mechanism that dynamically adjusts LVLMs' final outputs based on the token-level visual attention patterns, leveraging early-layer feedback to mitigate hallucinations caused by stylized images. Extensive experiments demonstrate that SAVER achieves state-of-the-art performance in hallucination mitigation across various models, datasets, and tasks.
Authors: Ulugbek Alibekov, Vanessa Staderini, Philipp Schneider, Doris Antensteiner
Abstract: This research focuses on visual industrial inspection by evaluating point clouds and multi-point cloud matching methods. We also introduce a synthetic dataset for quantitative evaluation of registration method and various distance metrics for point cloud comparison. Additionally, we present a novel CloudCompare plugin for merging multiple point clouds and visualizing surface defects, enhancing the accuracy and efficiency of automated inspection systems.
Authors: Weihang Liu, Yuke Li, Yuxuan Li, Jingyi Yu, Xin Lou
Abstract: Recent advances in 3D Gaussian Splatting (3DGS) have demonstrated remarkable rendering fidelity and efficiency. However, these methods still rely on computationally expensive sequential alpha-blending operations, resulting in significant overhead, particularly on resource-constrained platforms. In this paper, we propose Duplex-GS, a dual-hierarchy framework that integrates proxy Gaussian representations with order-independent rendering techniques to achieve photorealistic results while sustaining real-time performance. To mitigate the overhead caused by view-adaptive radix sort, we introduce cell proxies for local Gaussians management and propose cell search rasterization for further acceleration. By seamlessly combining our framework with Order-Independent Transparency (OIT), we develop a physically inspired weighted sum rendering technique that simultaneously eliminates "popping" and "transparency" artifacts, yielding substantial improvements in both accuracy and efficiency. Extensive experiments on a variety of real-world datasets demonstrate the robustness of our method across diverse scenarios, including multi-scale training views and large-scale environments. Our results validate the advantages of the OIT rendering paradigm in Gaussian Splatting, achieving high-quality rendering with an impressive 1.5 to 4 speedup over existing OIT based Gaussian Splatting approaches and 52.2% to 86.9% reduction of the radix sort overhead without quality degradation.
Authors: Heng Wu, Qian Zhang, Guixu Zhang
Abstract: Accurate monocular depth estimation remains a challenging problem due to the inherent ambiguity that stems from the ill-posed nature of recovering 3D structure from a single view, where multiple plausible depth configurations can produce identical 2D projections. In this paper, we present a novel depth estimation method that combines both local and global cues to improve prediction accuracy. Specifically, we propose the Gated Large Kernel Attention Module (GLKAM) to effectively capture multi-scale local structural information by leveraging large kernel convolutions with a gated mechanism. To further enhance the global perception of the network, we introduce the Global Bin Prediction Module (GBPM), which estimates the global distribution of depth bins and provides structural guidance for depth regression. Extensive experiments on the NYU-V2 and KITTI dataset demonstrate that our method achieves competitive performance and outperforms existing approaches, validating the effectiveness of each proposed component.
Authors: Tianshuo Zhang, Siran Peng, Li Gao, Haoyuan Zhang, Xiangyu Zhu, Zhen Lei
Abstract: The rapid advancements in face forgery techniques necessitate that detectors continuously adapt to new forgery methods, thus situating face forgery detection within a continual learning paradigm. However, when detectors learn new forgery types, their performance on previous types often degrades rapidly, a phenomenon known as catastrophic forgetting. Kolmogorov-Arnold Networks (KANs) utilize locally plastic splines as their activation functions, enabling them to learn new tasks by modifying only local regions of the functions while leaving other areas unaffected. Therefore, they are naturally suitable for addressing catastrophic forgetting. However, KANs have two significant limitations: 1) the splines are ineffective for modeling high-dimensional images, while alternative activation functions that are suitable for images lack the essential property of locality; 2) in continual learning, when features from different domains overlap, the mapping of different domains to distinct curve regions always collapses due to repeated modifications of the same regions. In this paper, we propose a KAN-based Continual Face Forgery Detection (KAN-CFD) framework, which includes a Domain-Group KAN Detector (DG-KD) and a data-free replay Feature Separation strategy via KAN Drift Compensation Projection (FS-KDCP). DG-KD enables KANs to fit high-dimensional image inputs while preserving locality and local plasticity. FS-KDCP avoids the overlap of the KAN input spaces without using data from prior tasks. Experimental results demonstrate that the proposed method achieves superior performance while notably reducing forgetting.
Authors: Tao Chen, Dan Zhang, Da Chen, Huazhu Fu, Kai Jin, Shanshan Wang, Laurent D. Cohen, Yitian Zhao, Quanyong Yi, Jiong Zhang
Abstract: Choroidal neovascularization (CNV), a primary characteristic of wet age-related macular degeneration (wet AMD), represents a leading cause of blindness worldwide. In clinical practice, optical coherence tomography angiography (OCTA) is commonly used for studying CNV-related pathological changes, due to its micron-level resolution and non-invasive nature. Thus, accurate segmentation of CNV regions and vessels in OCTA images is crucial for clinical assessment of wet AMD. However, challenges existed due to irregular CNV shapes and imaging limitations like projection artifacts, noises and boundary blurring. Moreover, the lack of publicly available datasets constraints the CNV analysis. To address these challenges, this paper constructs the first publicly accessible CNV dataset (CNVSeg), and proposes a novel multilateral graph convolutional interaction-enhanced CNV segmentation network (MTG-Net). This network integrates both region and vessel morphological information, exploring semantic and geometric duality constraints within the graph domain. Specifically, MTG-Net consists of a multi-task framework and two graph-based cross-task modules: Multilateral Interaction Graph Reasoning (MIGR) and Multilateral Reinforcement Graph Reasoning (MRGR). The multi-task framework encodes rich geometric features of lesion shapes and surfaces, decoupling the image into three task-specific feature maps. MIGR and MRGR iteratively reason about higher-order relationships across tasks through a graph mechanism, enabling complementary optimization for task-specific objectives. Additionally, an uncertainty-weighted loss is proposed to mitigate the impact of artifacts and noise on segmentation accuracy. Experimental results demonstrate that MTG-Net outperforms existing methods, achieving a Dice socre of 87.21\% for region segmentation and 88.12\% for vessel segmentation.
Authors: Yidan Wang, Chenyi Zhuang, Wutao Liu, Pan Gao, Nicu Sebe
Abstract: Weakly supervised visual grounding (VG) aims to locate objects in images based on text descriptions. Despite significant progress, existing methods lack strong cross-modal reasoning to distinguish subtle semantic differences in text expressions due to category-based and attribute-based ambiguity. To address these challenges, we introduce AlignCAT, a novel query-based semantic matching framework for weakly supervised VG. To enhance visual-linguistic alignment, we propose a coarse-grained alignment module that utilizes category information and global context, effectively mitigating interference from category-inconsistent objects. Subsequently, a fine-grained alignment module leverages descriptive information and captures word-level text features to achieve attribute consistency. By exploiting linguistic cues to their fullest extent, our proposed AlignCAT progressively filters out misaligned visual queries and enhances contrastive learning efficiency. Extensive experiments on three VG benchmarks, namely RefCOCO, RefCOCO+, and RefCOCOg, verify the superiority of AlignCAT against existing weakly supervised methods on two VG tasks. Our code is available at: https://github.com/I2-Multimedia-Lab/AlignCAT.
Authors: Ting Lei, Shaofeng Yin, Qingchao Chen, Yuxin Peng, Yang Liu
Abstract: Open Vocabulary Human-Object Interaction (HOI) detection aims to detect interactions between humans and objects while generalizing to novel interaction classes beyond the training set. Current methods often rely on Vision and Language Models (VLMs) but face challenges due to suboptimal image encoders, as image-level pre-training does not align well with the fine-grained region-level interaction detection required for HOI. Additionally, effectively encoding textual descriptions of visual appearances remains difficult, limiting the model's ability to capture detailed HOI relationships. To address these issues, we propose INteraction-aware Prompting with Concept Calibration (INP-CC), an end-to-end open-vocabulary HOI detector that integrates interaction-aware prompts and concept calibration. Specifically, we propose an interaction-aware prompt generator that dynamically generates a compact set of prompts based on the input scene, enabling selective sharing among similar interactions. This approach directs the model's attention to key interaction patterns rather than generic image-level semantics, enhancing HOI detection. Furthermore, we refine HOI concept representations through language model-guided calibration, which helps distinguish diverse HOI concepts by investigating visual similarities across categories. A negative sampling strategy is also employed to improve inter-modal similarity modeling, enabling the model to better differentiate visually similar but semantically distinct actions. Extensive experimental results demonstrate that INP-CC significantly outperforms state-of-the-art models on the SWIG-HOI and HICO-DET datasets. Code is available at https://github.com/ltttpku/INP-CC.
Authors: Xinwei Liu, Xiaojun Jia, Yuan Xun, Simeng Qin, Xiaochun Cao
Abstract: Vision-Language Models (VLMs) such as GPT-4o now demonstrate a remarkable ability to infer users' locations from public shared images, posing a substantial risk to geoprivacy. Although adversarial perturbations offer a potential defense, current methods are ill-suited for this scenario: they often perform poorly on high-resolution images and low perturbation budgets, and may introduce irrelevant semantic content. To address these limitations, we propose GeoShield, a novel adversarial framework designed for robust geoprivacy protection in real-world scenarios. GeoShield comprises three key modules: a feature disentanglement module that separates geographical and non-geographical information, an exposure element identification module that pinpoints geo-revealing regions within an image, and a scale-adaptive enhancement module that jointly optimizes perturbations at both global and local levels to ensure effectiveness across resolutions. Extensive experiments on challenging benchmarks show that GeoShield consistently surpasses prior methods in black-box settings, achieving strong privacy protection with minimal impact on visual or semantic quality. To our knowledge, this work is the first to explore adversarial perturbations for defending against geolocation inference by advanced VLMs, providing a practical and effective solution to escalating privacy concerns.
Authors: Wang Yu-Hang, Shiwei Li, Jianxiang Liao, Li Bohan, Jian Liu, Wenfei Yin
Abstract: Adversarial perturbations pose a significant threat to deep learning models. Adversarial Training (AT), the predominant defense method, faces challenges of high computational costs and a degradation in standard performance. While data augmentation offers an alternative path, existing techniques either yield limited robustness gains or incur substantial training overhead. Therefore, developing a defense mechanism that is both highly efficient and strongly robust is of paramount importance.In this work, we first conduct a systematic analysis of existing augmentation techniques, revealing that the synergy among diverse strategies -- rather than any single method -- is crucial for enhancing robustness. Based on this insight, we propose the Universal Adversarial Augmenter (UAA) framework, which is characterized by its plug-and-play nature and training efficiency. UAA decouples the expensive perturbation generation process from model training by pre-computing a universal transformation offline, which is then used to efficiently generate unique adversarial perturbations for each sample during training.Extensive experiments conducted on multiple benchmarks validate the effectiveness of UAA. The results demonstrate that UAA establishes a new state-of-the-art (SOTA) for data-augmentation-based adversarial defense strategies , without requiring the online generation of adversarial examples during training. This framework provides a practical and efficient pathway for building robust models,Our code is available in the supplementary materials.
Authors: Shanshan Guo, Xiwen Liang, Junfan Lin, Yuzheng Zhuang, Liang Lin, Xiaodan Liang
Abstract: Language-instructed robot manipulation has garnered significant interest due to the potential of learning from collected data. While the challenges in high-level perception and planning are continually addressed along the progress of general large pre-trained models, the low precision of low-level action estimation has emerged as the key limiting factor in manipulation performance. To this end, this paper introduces a novel robot manipulation framework, i.e., ActionSink, to pave the way toward precise action estimations in the field of learning-based robot manipulation. As the name suggests, ActionSink reformulates the actions of robots as action-caused optical flows from videos, called "action flow", in a self-supervised manner, which are then used to be retrieved and integrated to enhance the action estimation. Specifically, ActionSink incorporates two primary modules. The first module is a coarse-to-fine action flow matcher, which continuously refines the accuracy of action flow via iterative retrieval and denoising process. The second module is a dynamic action flow integrator, which employs a working memory pool that dynamically and efficiently manages the historical action flows that should be used to integrate to enhance the current action estimation. In this module, a multi-layer fusion module is proposed to integrate direct estimation and action flows from both the current and the working memory, achieving highly accurate action estimation through a series of estimation-integration processes. Our ActionSink framework outperformed prior SOTA on the LIBERO benchmark by a 7.9\% success rate, and obtained nearly an 8\% accuracy gain on the challenging long-horizon visual task LIBERO-Long.
Authors: Hongyu Shen, Junfeng Ni, Yixin Chen, Weishuo Li, Mingtao Pei, Siyuan Huang
Abstract: We address the challenge of lifting 2D visual segmentation to 3D in Gaussian Splatting. Existing methods often suffer from inconsistent 2D masks across viewpoints and produce noisy segmentation boundaries as they neglect these semantic cues to refine the learned Gaussians. To overcome this, we introduce Gaussian Instance Tracing (GIT), which augments the standard Gaussian representation with an instance weight matrix across input views. Leveraging the inherent consistency of Gaussians in 3D, we use this matrix to identify and correct 2D segmentation inconsistencies. Furthermore, since each Gaussian ideally corresponds to a single object, we propose a GIT-guided adaptive density control mechanism to split and prune ambiguous Gaussians during training, resulting in sharper and more coherent 2D and 3D segmentation boundaries. Experimental results show that our method extracts clean 3D assets and consistently improves 3D segmentation in both online (e.g., self-prompting) and offline (e.g., contrastive lifting) settings, enabling applications such as hierarchical segmentation, object extraction, and scene editing.
Authors: Freida Barnatan, Emunah Goldstein, Einav Kalimian, Orchen Madar, Avi Huri, David Zitoun, Ya'akov Mandelbaum, Moshe Amitay
Abstract: Accurate and efficient characterization of nanoparticle morphology in Scanning Electron Microscopy (SEM) images is critical for ensuring product quality in nanomaterial synthesis and accelerating development. However, conventional deep learning methods for shape classification require extensive labeled datasets and computationally demanding training, limiting their accessibility to the typical nanoparticle practitioner in research and industrial settings. In this study, we introduce a zero-shot classification pipeline that leverages two vision foundation models: the Segment Anything Model (SAM) for object segmentation and DINOv2 for feature embedding. By combining these models with a lightweight classifier, we achieve high-precision shape classification across three morphologically diverse nanoparticle datasets - without the need for extensive parameter fine-tuning. Our methodology outperforms a fine-tuned YOLOv11 and ChatGPT o4-mini-high baselines, demonstrating robustness to small datasets, subtle morphological variations, and domain shifts from natural to scientific imaging. Quantitative clustering metrics on PCA plots of the DINOv2 features are discussed as a means of assessing the progress of the chemical synthesis. This work highlights the potential of foundation models to advance automated microscopy image analysis, offering an alternative to traditional deep learning pipelines in nanoparticle research which is both more efficient and more accessible to the user.
Authors: Xingchao Yang, Shiori Ueda, Yuantian Huang, Tomoya Akiyama, Takafumi Taketomi
Abstract: Paired bare-makeup facial images are essential for a wide range of beauty-related tasks, such as virtual try-on, facial privacy protection, and facial aesthetics analysis. However, collecting high-quality paired makeup datasets remains a significant challenge. Real-world data acquisition is constrained by the difficulty of collecting large-scale paired images, while existing synthetic approaches often suffer from limited realism or inconsistencies between bare and makeup images. Current synthetic methods typically fall into two categories: warping-based transformations, which often distort facial geometry and compromise the precision of makeup; and text-to-image generation, which tends to alter facial identity and expression, undermining consistency. In this work, we present FFHQ-Makeup, a high-quality synthetic makeup dataset that pairs each identity with multiple makeup styles while preserving facial consistency in both identity and expression. Built upon the diverse FFHQ dataset, our pipeline transfers real-world makeup styles from existing datasets onto 18K identities by introducing an improved makeup transfer method that disentangles identity and makeup. Each identity is paired with 5 different makeup styles, resulting in a total of 90K high-quality bare-makeup image pairs. To the best of our knowledge, this is the first work that focuses specifically on constructing a makeup dataset. We hope that FFHQ-Makeup fills the gap of lacking high-quality bare-makeup paired datasets and serves as a valuable resource for future research in beauty-related tasks.
Authors: Lukas Ranftl, Felix Brendel, Bertram Drost, Carsten Steger
Abstract: We present MVTOP, a novel transformer-based method for multi-view rigid object pose estimation. Through an early fusion of the view-specific features, our method can resolve pose ambiguities that would be impossible to solve with a single view or with a post-processing of single-view poses. MVTOP models the multi-view geometry via lines of sight that emanate from the respective camera centers. While the method assumes the camera interior and relative orientations are known for a particular scene, they can vary for each inference. This makes the method versatile. The use of the lines of sight enables MVTOP to correctly predict the correct pose with the merged multi-view information. To show the model's capabilities, we provide a synthetic data set that can only be solved with such holistic multi-view approaches since the poses in the dataset cannot be solved with just one view. Our method outperforms single-view and all existing multi-view approaches on our dataset and achieves competitive results on the YCB-V dataset. To the best of our knowledge, no holistic multi-view method exists that can resolve such pose ambiguities reliably. Our model is end-to-end trainable and does not require any additional data, e.g., depth.
Authors: Chuanzhi Xu, Haoxian Zhou, Langyi Chen, Yuk Ying Chung, Qiang Qu
Abstract: Event cameras offer unparalleled advantages such as high temporal resolution, low latency, and high dynamic range. However, their limited spatial resolution poses challenges for fine-grained perception tasks. In this work, we propose an ultra-lightweight, stream-based event-to-event super-resolution method based on Spiking Neural Networks (SNNs), designed for real-time deployment on resource-constrained devices. To further reduce model size, we introduce a novel Dual-Forward Polarity-Split Event Encoding strategy that decouples positive and negative events into separate forward paths through a shared SNN. Furthermore, we propose a Learnable Spatio-temporal Polarity-aware Loss (LearnSTPLoss) that adaptively balances temporal, spatial, and polarity consistency using learnable uncertainty-based weights. Experimental results demonstrate that our method achieves competitive super-resolution performance on multiple datasets while significantly reducing model size and inference time. The lightweight design enables embedding the module into event cameras or using it as an efficient front-end preprocessing for downstream vision tasks.
Authors: Wentao Qu, Guofeng Mei, Jing Wang, Yujiao Wu, Xiaoshui Huang, Liang Xiao
Abstract: Denoising Diffusion Probabilistic Models (DDPMs) have shown success in robust 3D object detection tasks. Existing methods often rely on the score matching from 3D boxes or pre-trained diffusion priors. However, they typically require multi-step iterations in inference, which limits efficiency. To address this, we propose a \textbf{R}obust single-stage fully \textbf{S}parse 3D object \textbf{D}etection \textbf{Net}work with a Detachable Latent Framework (DLF) of DDPMs, named RSDNet. Specifically, RSDNet learns the denoising process in latent feature spaces through lightweight denoising networks like multi-level denoising autoencoders (DAEs). This enables RSDNet to effectively understand scene distributions under multi-level perturbations, achieving robust and reliable detection. Meanwhile, we reformulate the noising and denoising mechanisms of DDPMs, enabling DLF to construct multi-type and multi-level noise samples and targets, enhancing RSDNet robustness to multiple perturbations. Furthermore, a semantic-geometric conditional guidance is introduced to perceive the object boundaries and shapes, alleviating the center feature missing problem in sparse representations, enabling RSDNet to perform in a fully sparse detection pipeline. Moreover, the detachable denoising network design of DLF enables RSDNet to perform single-step detection in inference, further enhancing detection efficiency. Extensive experiments on public benchmarks show that RSDNet can outperform existing methods, achieving state-of-the-art detection.
Authors: Jisoo Kim, Wooseok Seo, Junwan Kim, Seungho Park, Sooyeon Park, Youngjae Yu
Abstract: With growing interest in deploying text-to-video (T2V) models in resource-constrained environments, reducing their high computational cost has become crucial, leading to extensive research on pruning and knowledge distillation methods while maintaining performance. However, existing distillation methods primarily rely on supervised fine-tuning (SFT), which often leads to mode collapse as pruned models with reduced capacity fail to directly match the teacher's outputs, ultimately resulting in degraded quality. To address this challenge, we propose an effective distillation method, ReDPO, that integrates DPO and SFT. Our approach leverages DPO to guide the student model to focus on recovering only the targeted properties, rather than passively imitating the teacher, while also utilizing SFT to enhance overall performance. We additionally propose V.I.P., a novel framework for filtering and curating high-quality pair datasets, along with a step-by-step online approach for calibrated training. We validate our method on two leading T2V models, VideoCrafter2 and AnimateDiff, achieving parameter reduction of 36.2% and 67.5% each, while maintaining or even surpassing the performance of full models. Further experiments demonstrate the effectiveness of both ReDPO and V.I.P. framework in enabling efficient and high-quality video generation. Our code and videos are available at https://jiiiisoo.github.io/VIP.github.io/.
Authors: Gang Dai, Yifan Zhang, Yutao Qin, Qiangya Guo, Shuangping Huang, Shuicheng Yan
Abstract: Existing handwritten text generation methods primarily focus on isolated words. However, realistic handwritten text demands attention not only to individual words but also to the relationships between them, such as vertical alignment and horizontal spacing. Therefore, generating entire text lines emerges as a more promising and comprehensive task. However, this task poses significant challenges, including the accurate modeling of complex style patterns encompassing both intra- and inter-word relationships, and maintaining content accuracy across numerous characters. To address these challenges, we propose DiffBrush, a novel diffusion-based model for handwritten text-line generation. Unlike existing methods, DiffBrush excels in both style imitation and content accuracy through two key strategies: (1) content-decoupled style learning, which disentangles style from content to better capture intra-word and inter-word style patterns by using column- and row-wise masking; and (2) multi-scale content learning, which employs line and word discriminators to ensure global coherence and local accuracy of textual content. Extensive experiments show that DiffBrush excels in generating high-quality text lines, particularly in style reproduction and content preservation. Code is available at https://github.com/dailenson/DiffBrush.
Authors: Huaihai Lyu, Chaofan Chen, Yuheng Ji, Changsheng Xu
Abstract: Driven by the increasing demand for applications in augmented and virtual reality, egocentric action recognition has emerged as a prominent research area. It is typically divided into two subtasks: recognizing the performed behavior (i.e., verb component) and identifying the objects being acted upon (i.e., noun component) from the first-person perspective. However, most existing approaches treat these two components as independent classification tasks, focusing on extracting component-specific knowledge while overlooking their inherent semantic and contextual relationships, leading to fragmented representations and sub-optimal generalization capability. To address these challenges, we propose a prompt learning-based framework, EgoPrompt, to conduct the egocentric action recognition task. Building on the existing prompting strategy to capture the component-specific knowledge, we construct a Unified Prompt Pool space to establish interaction between the two types of component representations. Specifically, the component representations (from verbs and nouns) are first decomposed into fine-grained patterns with the prompt pair form. Then, these pattern-level representations are fused through an attention-based mechanism to facilitate cross-component interaction. To ensure the prompt pool is informative, we further introduce a novel training objective, Diverse Pool Criteria. This objective realizes our goals from two perspectives: Prompt Selection Frequency Regularization and Prompt Knowledge Orthogonalization. Extensive experiments are conducted on the Ego4D, EPIC-Kitchens, and EGTEA datasets. The results consistently show that EgoPrompt achieves state-of-the-art performance across within-dataset, cross-dataset, and base-to-novel generalization benchmarks.
Authors: Hang Guo, Qing Zhang, Zixuan Gao, Siyuan Yang, Shulin Peng, Xiang Tao, Ting Yu, Yan Wang, Qingli Li
Abstract: Accurate prediction of placental diseases via whole slide images (WSIs) is critical for preventing severe maternal and fetal complications. However, WSI analysis presents significant computational challenges due to the massive data volume. Existing WSI classification methods encounter critical limitations: (1) inadequate patch selection strategies that either compromise performance or fail to sufficiently reduce computational demands, and (2) the loss of global histological context resulting from patch-level processing approaches. To address these challenges, we propose an Efficient multimodal framework for Patient-level placental disease Diagnosis, named EmmPD. Our approach introduces a two-stage patch selection module that combines parameter-free and learnable compression strategies, optimally balancing computational efficiency with critical feature preservation. Additionally, we develop a hybrid multimodal fusion module that leverages adaptive graph learning to enhance pathological feature representation and incorporates textual medical reports to enrich global contextual understanding. Extensive experiments conducted on both a self-constructed patient-level Placental dataset and two public datasets demonstrating that our method achieves state-of-the-art diagnostic performance. The code is available at https://github.com/ECNU-MultiDimLab/EmmPD.
Authors: Jun Luo, Zijing Zhao, Yang Liu
Abstract: Deep learning-based semantic segmentation models achieve impressive results yet remain limited in handling distribution shifts between training and test data. In this paper, we present SDGPA (Synthetic Data Generation and Progressive Adaptation), a novel method that tackles zero-shot domain adaptive semantic segmentation, in which no target images are available, but only a text description of the target domain's style is provided. To compensate for the lack of target domain training data, we utilize a pretrained off-the-shelf text-to-image diffusion model, which generates training images by transferring source domain images to target style. Directly editing source domain images introduces noise that harms segmentation because the layout of source images cannot be precisely maintained. To address inaccurate layouts in synthetic data, we propose a method that crops the source image, edits small patches individually, and then merges them back together, which helps improve spatial precision. Recognizing the large domain gap, SDGPA constructs an augmented intermediate domain, leveraging easier adaptation subtasks to enable more stable model adaptation to the target domain. Additionally, to mitigate the impact of noise in synthetic data, we design a progressive adaptation strategy, ensuring robust learning throughout the training process. Extensive experiments demonstrate that our method achieves state-of-the-art performance in zero-shot semantic segmentation. The code is available at https://github.com/ROUJINN/SDGPA
Authors: Libo Zhang, Xinyu Yi, Feng Xu
Abstract: In recent years, tracking human motion using IMUs from everyday devices such as smartphones and smartwatches has gained increasing popularity. However, due to the sparsity of sensor measurements and the lack of datasets capturing human motion over uneven terrain, existing methods often struggle with pose estimation accuracy and are typically limited to recovering movements on flat terrain only. To this end, we present BaroPoser, the first method that combines IMU and barometric data recorded by a smartphone and a smartwatch to estimate human pose and global translation in real time. By leveraging barometric readings, we estimate sensor height changes, which provide valuable cues for both improving the accuracy of human pose estimation and predicting global translation on non-flat terrain. Furthermore, we propose a local thigh coordinate frame to disentangle local and global motion input for better pose representation learning. We evaluate our method on both public benchmark datasets and real-world recordings. Quantitative and qualitative results demonstrate that our approach outperforms the state-of-the-art (SOTA) methods that use IMUs only with the same hardware configuration.
Authors: Mahdi Golizadeh, Nassibeh Golizadeh, Mohammad Ali Keyvanrad, Hossein Shirazi
Abstract: Object detection has achieved remarkable accuracy through deep learning, yet these improvements often come with increased computational cost, limiting deployment on resource-constrained devices. Knowledge Distillation (KD) provides an effective solution by enabling compact student models to learn from larger teacher models. However, adapting KD to object detection poses unique challenges due to its dual objectives-classification and localization-as well as foreground-background imbalance and multi-scale feature representation. This review introduces a novel architecture-centric taxonomy for KD methods, distinguishing between CNN-based detectors (covering backbone-level, neck-level, head-level, and RPN/RoI-level distillation) and Transformer-based detectors (including query-level, feature-level, and logit-level distillation). We further evaluate representative methods using the MS COCO and PASCAL VOC datasets with mAP@0.5 as performance metric, providing a comparative analysis of their effectiveness. The proposed taxonomy and analysis aim to clarify the evolving landscape of KD in object detection, highlight current challenges, and guide future research toward efficient and scalable detection systems.
Authors: Peiyu Wang, Yi Peng, Yimeng Gan, Liang Hu, Tianyidan Xie, Xiaokun Wang, Yichen Wei, Chuanxin Tang, Bo Zhu, Changshi Li, Hongyang Wei, Eric Li, Xuchen Song, Yang Liu, Yahui Zhou
Abstract: We introduce Skywork UniPic, a 1.5 billion-parameter autoregressive model that unifies image understanding, text-to-image generation, and image editing within a single architecture-eliminating the need for task-specific adapters or inter-module connectors-and demonstrate that compact multimodal systems can achieve state-of-the-art performance on commodity hardware. Skywork UniPic achieves a GenEval score of 0.86, surpassing most existing unified models; sets a new DPG-Bench complex-generation record of 85.5; attains 5.83 on GEditBench-EN and 3.49 on ImgEdit-Bench for image editing; and generates 1024 x 1024 images with under 15 GB of GPU memory (e.g., RTX 4090). (1) a decoupled encoding strategy that leverages a masked autoregressive encoder for synthesis and a SigLIP2 encoder for understanding, all feeding a shared autoregressive decoder; (2) a progressive, resolution-aware training schedule scaling from 256 x 256 to 1024 x 1024 while dynamically unfreezing parameters to balance capacity and stability; and (3) meticulously curated, 100 million-scale datasets augmented with task-specific reward models to refine generation and editing objectives. By demonstrating that high-fidelity multimodal integration need not incur prohibitive resource demands, Skywork UniPic establishes a practical paradigm for deployable, high-fidelity multimodal AI. Code and weights are publicly available at https://huggingface.co/Skywork/Skywork-UniPic-1.5B.
Authors: Satyapreet Singh Yadav, Akash K S, Chandra Sekhar Seelamantula, Chetan Singh Thakur
Abstract: We present a neuromorphic radar framework for real-time, low-power hand gesture recognition (HGR) using an event-driven architecture inspired by biological sensing. Our system comprises a 24 GHz Doppler radar front-end and a custom neuromorphic sampler that converts intermediate-frequency (IF) signals into sparse spike-based representations via asynchronous sigma-delta encoding. These events are directly processed by a lightweight neural network deployed on a Cortex-M0 microcontroller, enabling low-latency inference without requiring spectrogram reconstruction. Unlike conventional radar HGR pipelines that continuously sample and process data, our architecture activates only when meaningful motion is detected, significantly reducing memory, power, and computation overhead. Evaluated on a dataset of five gestures collected from seven users, our system achieves > 85% real-time accuracy. To the best of our knowledge, this is the first work that employs bio-inspired asynchronous sigma-delta encoding and an event-driven processing framework for radar-based HGR.
Authors: Amirreza Rouhi, Sneh Patel, Noah McCarthy, Siddiqa Khan, Hadi Khorsand, Kaleb Lefkowitz, David K. Han
Abstract: The exponential growth in Unmanned Aerial Vehicles (UAVs) usage underscores the critical need of detecting them at extended distances to ensure safe operations, especially in densely populated areas. Despite the tremendous advances made in computer vision through deep learning, the detection of these small airborne objects remains a formidable challenge. While several datasets have been developed specifically for drone detection, the need for a more extensive and diverse collection of drone image data persists, particularly for long-range detection under varying environmental conditions. We introduce here the Long Range Drone Detection (LRDD) Version 2 dataset, comprising 39,516 meticulously annotated images, as a second release of the LRDD dataset released previously. The LRDDv2 dataset enhances the LRDDv1 by incorporating a greater variety of images, providing a more diverse and comprehensive resource for drone detection research. What sets LRDDv2 apart is its inclusion of target range information for over 8,000 images, making it possible to develop algorithms for drone range estimation. Tailored for long-range aerial object detection, the majority of LRDDv2's dataset consists of images capturing drones with 50 or fewer pixels in 1080p resolution. For access to the complete Long-Range Drone Detection Dataset (LRDD)v2, please visit https://research.coe.drexel.edu/ece/imaple/lrddv2/ .
Authors: Xunzhi Xiang, Yabo Chen, Guiyu Zhang, Zhongyu Wang, Zhe Gao, Quanming Xiang, Gonghu Shang, Junqi Liu, Haibin Huang, Yang Gao, Chi Zhang, Qi Fan, Xuelong Li
Abstract: Current autoregressive diffusion models excel at video generation but are generally limited to short temporal durations. Our theoretical analysis indicates that the autoregressive modeling typically suffers from temporal drift caused by error accumulation and hinders parallelization in long video synthesis. To address these limitations, we propose a novel planning-then-populating framework centered on Macro-from-Micro Planning (MMPL) for long video generation. MMPL sketches a global storyline for the entire video through two hierarchical stages: Micro Planning and Macro Planning. Specifically, Micro Planning predicts a sparse set of future keyframes within each short video segment, offering motion and appearance priors to guide high-quality video segment generation. Macro Planning extends the in-segment keyframes planning across the entire video through an autoregressive chain of micro plans, ensuring long-term consistency across video segments. Subsequently, MMPL-based Content Populating generates all intermediate frames in parallel across segments, enabling efficient parallelization of autoregressive generation. The parallelization is further optimized by Adaptive Workload Scheduling for balanced GPU execution and accelerated autoregressive video generation. Extensive experiments confirm that our method outperforms existing long video generation models in quality and stability. Generated videos and comparison results are in our project page.
Authors: Tongshun Zhang, Pingping Liu, Zixuan Zhong, Zijian Zhang, Qiuzhan Zhou
Abstract: Recovering fine-grained details in extremely dark images remains challenging due to severe structural information loss and noise corruption. Existing enhancement methods often fail to preserve intricate details and sharp edges, limiting their effectiveness in downstream applications like text and edge detection. To address these deficiencies, we propose an efficient dual-stage approach centered on detail recovery for dark images. In the first stage, we introduce a Residual Fourier-Guided Module (RFGM) that effectively restores global illumination in the frequency domain. RFGM captures inter-stage and inter-channel dependencies through residual connections, providing robust priors for high-fidelity frequency processing while mitigating error accumulation risks from unreliable priors. The second stage employs complementary Mamba modules specifically designed for textural structure refinement: (1) Patch Mamba operates on channel-concatenated non-downsampled patches, meticulously modeling pixel-level correlations to enhance fine-grained details without resolution loss. (2) Grad Mamba explicitly focuses on high-gradient regions, alleviating state decay in state space models and prioritizing reconstruction of sharp edges and boundaries. Extensive experiments on multiple benchmark datasets and downstream applications demonstrate that our method significantly improves detail recovery performance while maintaining efficiency. Crucially, the proposed modules are lightweight and can be seamlessly integrated into existing Fourier-based frameworks with minimal computational overhead. Code is available at https://github.com/bywlzts/RFGM.
Authors: Shaoguang Wang, Jianxiang He, Yijie Xu, Ziyang Chen, Weiyu Guo, Hui Xiong
Abstract: The practical application of Multimodal Large Language Models (MLLMs) to Video Question Answering (Video-QA) is severely hindered by the high token cost of processing numerous video frames. While increasing the number of sampled frames is a common strategy, we observe a "less is more" phenomenon where excessive frames can paradoxically degrade performance due to context dilution. Concurrently, state-of-the-art keyframe selection methods, while effective, still yield significant temporal redundancy, which we term 'visual echoes'. To address these dual challenges, we propose Adaptive Frame-Pruning (AFP), a novel post-processing method that intelligently prunes the selected keyframes. AFP employs an adaptive hierarchical clustering algorithm on a fused ResNet-50 and CLIP feature space to identify and merge these echoes into single representatives. To compensate for information loss, we then introduce a lightweight, text-based semantic graph that provides critical context with minimal token overhead. Conducting extensive experiments on the LongVideoBench and VideoMME benchmarks across multiple leading MLLMs, our full approach demonstrates a drastic reduction in required frames by up to 86.9% and total input tokens by up to 83.2%. Crucially, by providing a concise, high-quality set of frames, our method not only enhances efficiency but often improves accuracy over baselines that use more frames. The code will be released upon publication.
Authors: Tongshun Zhang, Pingping Liu, Zhe Zhang, Qiuzhan Zhou
Abstract: Images captured in nighttime scenes suffer from severely reduced visibility, hindering effective content perception. Current low-light image enhancement (LLIE) methods face significant challenges: data-driven end-to-end mapping networks lack interpretability or rely on unreliable prior guidance, struggling under extremely dark conditions, while physics-based methods depend on simplified assumptions that often fail in complex real-world scenarios. To address these limitations, we propose CIVQLLIE, a novel framework that leverages the power of discrete representation learning through causal reasoning. We achieve this through Vector Quantization (VQ), which maps continuous image features to a discrete codebook of visual tokens learned from large-scale high-quality images. This codebook serves as a reliable prior, encoding standardized brightness and color patterns that are independent of degradation. However, direct application of VQ to low-light images fails due to distribution shifts between degraded inputs and the learned codebook. Therefore, we propose a multi-level causal intervention approach to systematically correct these shifts. First, during encoding, our Pixel-level Causal Intervention (PCI) module intervenes to align low-level features with the brightness and color distributions expected by the codebook. Second, a Feature-aware Causal Intervention (FCI) mechanism with Low-frequency Selective Attention Gating (LSAG) identifies and enhances channels most affected by illumination degradation, facilitating accurate codebook token matching while enhancing the encoder's generalization performance through flexible feature-level intervention. Finally, during decoding, the High-frequency Detail Reconstruction Module (HDRM) leverages structural information preserved in the matched codebook representations to reconstruct fine details using deformable convolution techniques.
Authors: Junlong Ren, Gangjian Zhang, Honghao Fu, Pengcheng Wu, Hao Wang
Abstract: Text-Motion Retrieval (TMR) aims to retrieve 3D motion sequences semantically relevant to text descriptions. However, matching 3D motions with text remains highly challenging, primarily due to the intricate structure of human body and its spatial-temporal dynamics. Existing approaches often overlook these complexities, relying on general encoding methods that fail to distinguish different body parts and their dynamics, limiting precise semantic alignment. To address this, we propose WaMo, a novel wavelet-based multi-frequency feature extraction framework. It fully captures part-specific and time-varying motion details across multiple resolutions on body joints, extracting discriminative motion features to achieve fine-grained alignment with texts. WaMo has three key components: (1) Trajectory Wavelet Decomposition decomposes motion signals into frequency components that preserve both local kinematic details and global motion semantics. (2) Trajectory Wavelet Reconstruction uses learnable inverse wavelet transforms to reconstruct original joint trajectories from extracted features, ensuring the preservation of essential spatial-temporal information. (3) Disordered Motion Sequence Prediction reorders shuffled motion sequences to improve the learning of inherent temporal coherence, enhancing motion-text alignment. Extensive experiments demonstrate WaMo's superiority, achieving 17.0\% and 18.2\% improvements in $Rsum$ on HumanML3D and KIT-ML datasets, respectively, outperforming existing state-of-the-art (SOTA) methods.
Authors: Yufei Xue, Yushi Huang, Jiawei Shao, Jun Zhang
Abstract: Post-training quantization (PTQ) has emerged as an effective approach for compressing large models and accelerating their inference without retraining. While PTQ has been extensively studied in the context of large language models (LLMs), its applicability to vision-language models (VLMs) remains underexplored. In this paper, we identify a modality discrepancy (\emph{i.e.}, limited text tokens \emph{vs.} excessive and redundant vision tokens) of VLMs. However, existing Hessian-based LLM PTQ methods treat all tokens equally during quantization, resulting in severe performance drops when applied to VLMs. Motivated by this observation, we propose a novel importance-aware PTQ framework tailored for VLMs, dubbed VLMQ. Specifically, to address vision token redundancy, VLMQ 1) optimizes an importance-aware objective that yields an enhanced Hessian with token-level importance factors, while retaining compatibility with parallelized weight updates, and 2) ensures efficiency and effectiveness by computing these factors via a single lightweight block-wise backward pass, guided by a theoretical connection to token-level perturbations. Extensive evaluations on 8 benchmarks across 0.5B$\sim$32B VLMs demonstrate the state-of-the-art (SOTA) performance of our VLMQ, particularly under low-bit settings. For example, it achieves a substantial \textbf{16.45\%} improvement on MME-RealWorld under 2-bit quantization.
Authors: Matteo Caligiuri, Francesco Barbato, Donald Shenaj, Umberto Michieli, Pietro Zanuttigh
Abstract: Federated Learning (FL) is an established paradigm for training deep learning models on decentralized data. However, as the size of the models grows, conventional FL approaches often require significant computational resources on client devices, which may not be feasible. We introduce FedPromo, a novel framework that enables efficient adaptation of large-scale foundation models stored on a central server to new domains encountered only by remote clients. Instead of directly training the large model on client devices, FedPromo optimizes lightweight proxy models via FL, significantly reducing computational overhead while maintaining privacy. Our method follows a two-stage process: first, server-side knowledge distillation aligns the representations of a large-scale foundation model (e.g., a transformer) with those of a compact counterpart (e.g., a CNN). Then, the compact model encoder is deployed to client devices, where trainable classifiers are learned locally. These classifiers are subsequently aggregated and seamlessly transferred back to the foundation model, facilitating personalized adaptation without requiring direct access to user data. Through novel regularization strategies, our framework enables decentralized multi-domain learning, balancing performance, privacy, and resource efficiency. Extensive experiments on five image classification benchmarks demonstrate that FedPromo outperforms existing methods while assuming limited-resource clients.
Authors: Ni Tang, Xiaotong Luo, Zihan Cheng, Liangtai Zhou, Dongxiao Zhang, Yanyun Qu
Abstract: Diffusion models have revealed powerful potential in all-in-one image restoration (AiOIR), which is talented in generating abundant texture details. The existing AiOIR methods either retrain a diffusion model or fine-tune the pretrained diffusion model with extra conditional guidance. However, they often suffer from high inference costs and limited adaptability to diverse degradation types. In this paper, we propose an efficient AiOIR method, Diffusion Once and Done (DOD), which aims to achieve superior restoration performance with only one-step sampling of Stable Diffusion (SD) models. Specifically, multi-degradation feature modulation is first introduced to capture different degradation prompts with a pretrained diffusion model. Then, parameter-efficient conditional low-rank adaptation integrates the prompts to enable the fine-tuning of the SD model for adapting to different degradation types. Besides, a high-fidelity detail enhancement module is integrated into the decoder of SD to improve structural and textural details. Experiments demonstrate that our method outperforms existing diffusion-based restoration approaches in both visual quality and inference efficiency.
Authors: Keyi Li, Alexander Jaus, Jens Kleesiek, Rainer Stiefelhagen
Abstract: Radiologists rely on anatomical understanding to accurately delineate pathologies, yet most current deep learning approaches use pure pattern recognition and ignore the anatomical context in which pathologies develop. To narrow this gap, we introduce GRASP (Guided Representation Alignment for the Segmentation of Pathologies), a modular plug-and-play framework that enhances pathology segmentation models by leveraging existing anatomy segmentation models through pseudolabel integration and feature alignment. Unlike previous approaches that obtain anatomical knowledge via auxiliary training, GRASP integrates into standard pathology optimization regimes without retraining anatomical components. We evaluate GRASP on two PET/CT datasets, conduct systematic ablation studies, and investigate the framework's inner workings. We find that GRASP consistently achieves top rankings across multiple evaluation metrics and diverse architectures. The framework's dual anatomy injection strategy, combining anatomical pseudo-labels as input channels with transformer-guided anatomical feature fusion, effectively incorporates anatomical context.
Authors: Jingjie Wang, Shunli Zhang, Xiang Wei, Senmao Tian
Abstract: Current gait recognition methodologies generally necessitate retraining when encountering new datasets. Nevertheless, retrained models frequently encounter difficulties in preserving knowledge from previous datasets, leading to a significant decline in performance on earlier test sets. To tackle these challenges, we present a continual gait recognition task, termed GaitAdapt, which supports the progressive enhancement of gait recognition capabilities over time and is systematically categorized according to various evaluation scenarios. Additionally, we propose GaitAdapter, a non-replay continual learning approach for gait recognition. This approach integrates the GaitPartition Adaptive Knowledge (GPAK) module, employing graph neural networks to aggregate common gait patterns from current data into a repository constructed from graph vectors. Subsequently, this repository is used to improve the discriminability of gait features in new tasks, thereby enhancing the model's ability to effectively recognize gait patterns. We also introduce a Euclidean Distance Stability Method (EDSN) based on negative pairs, which ensures that newly added gait samples from different classes maintain similar relative spatial distributions across both previous and current gait tasks, thereby alleviating the impact of task changes on the distinguishability of original domain features. Extensive evaluations demonstrate that GaitAdapter effectively retains gait knowledge acquired from diverse tasks, exhibiting markedly superior discriminative capability compared to alternative methods.
Authors: Yizhe Xiong, Zihan Zhou, Yiwen Liang, Hui Chen, Zijia Lin, Tianxiang Hao, Fan Zhang, Jungong Han, Guiguang Ding
Abstract: Test-Time Adaptation (TTA) has emerged as an effective solution for adapting Vision Transformers (ViT) to distribution shifts without additional training data. However, existing TTA methods often incur substantial computational overhead, limiting their applicability in resource-constrained real-world scenarios. To reduce inference cost, plug-and-play token aggregation methods merge redundant tokens in ViTs to reduce total processed tokens. Albeit efficient, it suffers from significant performance degradation when directly integrated with existing TTA methods. We formalize this problem as Efficient Test-Time Adaptation (ETTA), seeking to preserve the adaptation capability of TTA while reducing inference latency. In this paper, we first provide a theoretical analysis from a novel mutual information perspective, showing that token aggregation inherently leads to information loss, which cannot be fully mitigated by conventional norm-tuning-based TTA methods. Guided by this insight, we propose to \textbf{N}eutralize Token \textbf{A}ggregation \textbf{v}ia \textbf{I}nformation \textbf{A}ugmentation (\textbf{NAVIA}). Specifically, we directly augment the [CLS] token embedding and incorporate adaptive biases into the [CLS] token in shallow layers of ViTs. We theoretically demonstrate that these augmentations, when optimized via entropy minimization, recover the information lost due to token aggregation. Extensive experiments across various out-of-distribution benchmarks demonstrate that NAVIA significantly outperforms state-of-the-art methods by over 2.5\%, while achieving an inference latency reduction of more than 20\%, effectively addressing the ETTA challenge.
Authors: Xinzhu Li, Juepeng Zheng, Yikun Chen, Xudong Mao, Guanghui Yue, Wei Zhou, Chenlei Lv, Ruomei Wang, Fan Zhou, Baoquan Zhao
Abstract: Robust gait recognition requires highly discriminative representations, which are closely tied to input modalities. While binary silhouettes and skeletons have dominated recent literature, these 2D representations fall short of capturing sufficient cues that can be exploited to handle viewpoint variations, and capture finer and meaningful details of gait. In this paper, we introduce a novel framework, termed DepthGait, that incorporates RGB-derived depth maps and silhouettes for enhanced gait recognition. Specifically, apart from the 2D silhouette representation of the human body, the proposed pipeline explicitly estimates depth maps from a given RGB image sequence and uses them as a new modality to capture discriminative features inherent in human locomotion. In addition, a novel multi-scale and cross-level fusion scheme has also been developed to bridge the modality gap between depth maps and silhouettes. Extensive experiments on standard benchmarks demonstrate that the proposed DepthGait achieves state-of-the-art performance compared to peer methods and attains an impressive mean rank-1 accuracy on the challenging datasets.
Authors: Pingchuan Ma, Xiaopei Yang, Yusong Li, Ming Gui, Felix Krause, Johannes Schusterbauer, Bj\"orn Ommer
Abstract: Explicitly disentangling style and content in vision models remains challenging due to their semantic overlap and the subjectivity of human perception. Existing methods propose separation through generative or discriminative objectives, but they still face the inherent ambiguity of disentangling intertwined concepts. Instead, we ask: Can we bypass explicit disentanglement by learning to merge style and content invertibly, allowing separation to emerge naturally? We propose SCFlow, a flow-matching framework that learns bidirectional mappings between entangled and disentangled representations. Our approach is built upon three key insights: 1) Training solely to merge style and content, a well-defined task, enables invertible disentanglement without explicit supervision; 2) flow matching bridges on arbitrary distributions, avoiding the restrictive Gaussian priors of diffusion models and normalizing flows; and 3) a synthetic dataset of 510,000 samples (51 styles $\times$ 10,000 content samples) was curated to simulate disentanglement through systematic style-content pairing. Beyond controllable generation tasks, we demonstrate that SCFlow generalizes to ImageNet-1k and WikiArt in zero-shot settings and achieves competitive performance, highlighting that disentanglement naturally emerges from the invertible merging process.
Authors: Gang Yang
Abstract: Hyperspectral unmixing aims at estimating material signatures (known as endmembers) and the corresponding proportions (referred to abundances), which is a critical preprocessing step in various hyperspectral imagery applications. This study develops a novel approach called sparsity and total variation (TV) constrained multilayer linear unmixing (STVMLU) for hyperspectral imagery. Specifically, based on a multilayer matrix factorization model, to improve the accuracy of unmixing, a TV constraint is incorporated to consider adjacent spatial similarity. Additionally, a L1/2-norm sparse constraint is adopted to effectively characterize the sparsity of the abundance matrix. For optimizing the STVMLU model, the method of alternating direction method of multipliers (ADMM) is employed, which allows for the simultaneous extraction of endmembers and their corresponding abundance matrix. Experimental results illustrate the enhanced performance of the proposed STVMLU when compared to other algorithms.
Authors: Xinlei Yu, Zhangquan Chen, Yudong Zhang, Shilin Lu, Ruolin Shen, Jiangning Zhang, Xiaobin Hu, Yanwei Fu, Shuicheng Yan
Abstract: Existing vision-language models (VLMs), whether generalists or specialists, remain constrained by their parameter scale, lack robust self-correction capabilities, and underperform in tasks involving long visual contexts and complex reasoning, resulting in suboptimal performance on document-based tasks. To address this, we propose MACT, a Multi-Agent Collaboration framework with Test-Time scaling, tailored for visual document understanding and visual question answering (VQA). It comprises four distinct small-scale agents, i.e., planning, execution, judgment, and answer agents, with clearly defined roles and effective collaboration. Notably, the judgment agent exclusively verifies correctness and redirects to prior agents for revisions, outperforming conventional correction strategies. To further expand the capability boundaries of the framework, we propose mixed reward modeling that balances agent-specific abilities and global collaboration, as well as agent-wise hybrid test-time scaling, which customizes different scaling strategies for each agent based on their functions. Evaluated on benchmarks spanning both document-based and non-document-based settings, our MACT shows superior performance with a smaller parameter scale without sacrificing the ability of general and mathematical tasks. Especially, it stands out in benchmarks involving long visual contexts and complicated reasoning. The three variants of MACT consistently hold the top three positions in average scores, leading in 13 of the 15 benchmarks. Code will be available at: https://github.com/YU-deep/MACT.git.
Authors: Diana-Nicoleta Grigore, Neelu Madan, Andreas Mogelmose, Thomas B. Moeslund, Radu Tudor Ionescu
Abstract: Unsupervised video segmentation is a challenging computer vision task, especially due to the lack of supervisory signals coupled with the complexity of visual scenes. To overcome this challenge, state-of-the-art models based on slot attention often have to rely on large and computationally expensive neural architectures. To this end, we propose a simple knowledge distillation framework that effectively transfers object-centric representations to a lightweight student. The proposed framework, called SlotMatch, aligns corresponding teacher and student slots via the cosine similarity, requiring no additional distillation objectives or auxiliary supervision. The simplicity of SlotMatch is confirmed via theoretical and empirical evidence, both indicating that integrating additional losses is redundant. We conduct experiments on two datasets to compare the state-of-the-art teacher model, SlotContrast, with our distilled student. The results show that our student based on SlotMatch matches and even outperforms its teacher, while using 3.6x less parameters and running 1.9x faster. Moreover, our student surpasses previous unsupervised video segmentation models.
Authors: Shivangi Nigam, Adarsh Prasad Behera, Shekhar Verma, P. Nagabhushan
Abstract: This paper presents Fd-CycleGAN, an image-to-image (I2I) translation framework that enhances latent representation learning to approximate real data distributions. Building upon the foundation of CycleGAN, our approach integrates Local Neighborhood Encoding (LNE) and frequency-aware supervision to capture fine-grained local pixel semantics while preserving structural coherence from the source domain. We employ distribution-based loss metrics, including KL/JS divergence and log-based similarity measures, to explicitly quantify the alignment between real and generated image distributions in both spatial and frequency domains. To validate the efficacy of Fd-CycleGAN, we conduct experiments on diverse datasets -- Horse2Zebra, Monet2Photo, and a synthetically augmented Strike-off dataset. Compared to baseline CycleGAN and other state-of-the-art methods, our approach demonstrates superior perceptual quality, faster convergence, and improved mode diversity, particularly in low-data regimes. By effectively capturing local and global distribution characteristics, Fd-CycleGAN achieves more visually coherent and semantically consistent translations. Our results suggest that frequency-guided latent learning significantly improves generalization in image translation tasks, with promising applications in document restoration, artistic style transfer, and medical image synthesis. We also provide comparative insights with diffusion-based generative models, highlighting the advantages of our lightweight adversarial approach in terms of training efficiency and qualitative output.
Authors: Futian Wang, Yuhan Qiao, Xiao Wang, Fuling Wang, Yuxiang Zhang, Dengdi Sun
Abstract: X-ray medical report generation is one of the important applications of artificial intelligence in healthcare. With the support of large foundation models, the quality of medical report generation has significantly improved. However, challenges such as hallucination and weak disease diagnostic capability still persist. In this paper, we first construct a large-scale multi-modal medical knowledge graph (termed M3KG) based on the ground truth medical report using the GPT-4o. It contains 2477 entities, 3 kinds of relations, 37424 triples, and 6943 disease-aware vision tokens for the CheXpert Plus dataset. Then, we sample it to obtain multi-granularity semantic graphs and use an R-GCN encoder for feature extraction. For the input X-ray image, we adopt the Swin-Transformer to extract the vision features and interact with the knowledge using cross-attention. The vision tokens are fed into a Q-former and retrieved the disease-aware vision tokens using another cross-attention. Finally, we adopt the large language model to map the semantic knowledge graph, input X-ray image, and disease-aware vision tokens into language descriptions. Extensive experiments on multiple datasets fully validated the effectiveness of our proposed knowledge graph and X-ray report generation framework. The source code of this paper will be released on https://github.com/Event-AHU/Medical_Image_Analysis.
Authors: Hongjun Liu, Chao Yao, Yalan Zhang, Xiaokun wang, Xiaojuan Ban
Abstract: Electroencephalogram (EEG) signal classification faces significant challenges due to data distribution shifts caused by heterogeneous electrode configurations, acquisition protocols, and hardware discrepancies across domains. This paper introduces IMAC, a novel channel-dependent mask and imputation self-supervised framework that formulates the alignment of cross-domain EEG data shifts as a spatial time series imputation task. To address heterogeneous electrode configurations in cross-domain scenarios, IMAC first standardizes different electrode layouts using a 3D-to-2D positional unification mapping strategy, establishing unified spatial representations. Unlike previous mask-based self-supervised representation learning methods, IMAC introduces spatio-temporal signal alignment. This involves constructing a channel-dependent mask and reconstruction task framed as a low-to-high resolution EEG spatial imputation problem. Consequently, this approach simulates cross-domain variations such as channel omissions and temporal instabilities, thus enabling the model to leverage the proposed imputer for robust signal alignment during inference. Furthermore, IMAC incorporates a disentangled structure that separately models the temporal and spatial information of the EEG signals separately, reducing computational complexity while enhancing flexibility and adaptability. Comprehensive evaluations across 10 publicly available EEG datasets demonstrate IMAC's superior performance, achieving state-of-the-art classification accuracy in both cross-subject and cross-center validation scenarios. Notably, IMAC shows strong robustness under both simulated and real-world distribution shifts, surpassing baseline methods by up to $35$\% in integrity scores while maintaining consistent classification accuracy.
Authors: Ning Zhu, Xiaochuan Ma, Shaoting Zhang, Guotai Wang
Abstract: Cold-Start Active Learning (CSAL) aims to select informative samples for annotation without prior knowledge, which is important for improving annotation efficiency and model performance under a limited annotation budget in medical image analysis. Most existing CSAL methods rely on Self-Supervised Learning (SSL) on the target dataset for feature extraction, which is inefficient and limited by insufficient feature representation. Recently, pre-trained Foundation Models (FMs) have shown powerful feature extraction ability with a potential for better CSAL. However, this paradigm has been rarely investigated, with a lack of benchmarks for comparison of FMs in CSAL tasks. To this end, we propose MedCAL-Bench, the first systematic FM-based CSAL benchmark for medical image analysis. We evaluate 14 FMs and 7 CSAL strategies across 7 datasets under different annotation budgets, covering classification and segmentation tasks from diverse medical modalities. It is also the first CSAL benchmark that evaluates both the feature extraction and sample selection stages. Our experimental results reveal that: 1) Most FMs are effective feature extractors for CSAL, with DINO family performing the best in segmentation; 2) The performance differences of these FMs are large in segmentation tasks, while small for classification; 3) Different sample selection strategies should be considered in CSAL on different datasets, with Active Learning by Processing Surprisal (ALPS) performing the best in segmentation while RepDiv leading for classification. The code is available at https://github.com/HiLab-git/MedCAL-Bench.
Authors: Shangwen Zhu, Qianyu Peng, Yuting Hu, Zhantao Yang, Han Zhang, Zhao Pu, Ruili Feng, Fan Cheng
Abstract: Flow-based generative models have recently achieved remarkable progress in image and video synthesis, with classifier-free guidance (CFG) becoming the standard tool for high-fidelity, controllable generation. However, despite their practical success, little is known about how guidance interacts with different stages of the sampling process-especially in the fast, low-step regimes typical of modern flow-based pipelines. In this work, we uncover and analyze a fundamental instability: the earliest reverse steps are acutely sensitive to the guidance scale, owing to a pronounced spike in the relative strength (RATIO) of conditional to unconditional predictions. Through rigorous theoretical analysis and empirical validation, we show that this RATIO spike is intrinsic to the data distribution, independent of the model architecture, and causes exponential error amplification when paired with strong guidance. To address this, we propose a simple, theoretically grounded, RATIO-aware adaptive guidance schedule that automatically dampens the guidance scale at early steps based on the evolving RATIO, using a closed-form exponential decay. Our method is lightweight, requires no additional inference overhead, and is compatible with standard flow frameworks. Experiments across state-of-the-art image (SD3.5, Lumina) and video (WAN2.1) models demonstrate that our approach enables up to 3x faster sampling while maintaining or improving generation quality, robustness, and semantic alignment. Extensive ablation studies further confirm the generality and stability of our schedule across models, datasets, and hyperparameters. Our findings highlight the critical role of stepwise guidance adaptation in unlocking the full potential of fast flow-based generative models.
Authors: Qiyu Chen, Zhen Qu, Wei Luo, Haiming Yao, Yunkang Cao, Yuxin Jiang, Yinan Duan, Huiyuan Luo, Chengkan Lv, Zhengtao Zhang
Abstract: Recently, large pre-trained vision-language models have shown remarkable performance in zero-shot anomaly detection (ZSAD). With fine-tuning on a single auxiliary dataset, the model enables cross-category anomaly detection on diverse datasets covering industrial defects and medical lesions. Compared to manually designed prompts, prompt learning eliminates the need for expert knowledge and trial-and-error. However, it still faces the following challenges: (i) static learnable tokens struggle to capture the continuous and diverse patterns of normal and anomalous states, limiting generalization to unseen categories; (ii) fixed textual labels provide overly sparse category information, making the model prone to overfitting to a specific semantic subspace. To address these issues, we propose Conditional Prompt Synthesis (CoPS), a novel framework that synthesizes dynamic prompts conditioned on visual features to enhance ZSAD performance. Specifically, we extract representative normal and anomaly prototypes from fine-grained patch features and explicitly inject them into prompts, enabling adaptive state modeling. Given the sparsity of class labels, we leverage a variational autoencoder to model semantic image features and implicitly fuse varied class tokens into prompts. Additionally, integrated with our spatially-aware alignment mechanism, extensive experiments demonstrate that CoPS surpasses state-of-the-art methods by 2.5% AUROC in both classification and segmentation across 13 industrial and medical datasets. Code will be available at https://github.com/cqylunlun/CoPS.
Authors: Xuan Dong, Xiangyuan Sun, Xia Wang, Jian Song, Ya Li, Weixin Li
Abstract: Moire patterns, unwanted color artifacts in images and videos, arise from the interference between spatially high-frequency scene contents and the spatial discrete sampling of digital cameras. Existing demoireing methods primarily rely on single-camera image/video processing, which faces two critical challenges: 1) distinguishing moire patterns from visually similar real textures, and 2) preserving tonal consistency and temporal coherence while removing moire artifacts. To address these issues, we propose a dual-camera framework that captures synchronized videos of the same scene: one in focus (retaining high-quality textures but may exhibit moire patterns) and one defocused (with significantly reduced moire patterns but blurred textures). We use the defocused video to help distinguish moire patterns from real texture, so as to guide the demoireing of the focused video. We propose a frame-wise demoireing pipeline, which begins with an optical flow based alignment step to address any discrepancies in displacement and occlusion between the focused and defocused frames. Then, we leverage the aligned defocused frame to guide the demoireing of the focused frame using a multi-scale CNN and a multi-dimensional training loss. To maintain tonal and temporal consistency, our final step involves a joint bilateral filter to leverage the demoireing result from the CNN as the guide to filter the input focused frame to obtain the final output. Experimental results demonstrate that our proposed framework largely outperforms state-of-the-art image and video demoireing methods.
Authors: Zilin Chen, Shengnan Lu
Abstract: Accurate detection of polyps is of critical importance for the early and intermediate stages of colorectal cancer diagnosis. Compared to static images, dynamic colonoscopy videos provide more comprehensive visual information, which can facilitate the development of effective treatment plans. However, unlike fixed-camera recordings, colonoscopy videos often exhibit rapid camera movement, introducing substantial background noise that disrupts the structural integrity of the scene and increases the risk of false positives. To address these challenges, we propose the Adaptive Video Polyp Detection Network (AVPDN), a robust framework for multi-scale polyp detection in colonoscopy videos. AVPDN incorporates two key components: the Adaptive Feature Interaction and Augmentation (AFIA) module and the Scale-Aware Context Integration (SACI) module. The AFIA module adopts a triple-branch architecture to enhance feature representation. It employs dense self-attention for global context modeling, sparse self-attention to mitigate the influence of low query-key similarity in feature aggregation, and channel shuffle operations to facilitate inter-branch information exchange. In parallel, the SACI module is designed to strengthen multi-scale feature integration. It utilizes dilated convolutions with varying receptive fields to capture contextual information at multiple spatial scales, thereby improving the model's denoising capability. Experiments conducted on several challenging public benchmarks demonstrate the effectiveness and generalization ability of the proposed method, achieving competitive performance in video-based polyp detection tasks.
Authors: Jiabing Yang, Chenhang Cui, Yiyang Zhou, Yixiang Chen, Peng Xia, Ying Wei, Tao Yu, Yan Huang, Liang Wang
Abstract: Recent advancements in Large Vision-Language Models (LVLMs) have demonstrated significant progress across multiple domains. However, these models still face the inherent challenge of integrating vision and language for collaborative inference, which often leads to "hallucinations", outputs that are not grounded in the corresponding images. Many efforts have been made to address these issues, but each comes with its own limitations, such as high computational cost or expensive dataset annotation. Recent research shows that LVLMs exhibit a long-term bias where hallucinations increase as the sequence length grows, yet the underlying cause remains poorly understood. Building on extensive research into attention mechanisms in LVLMs, we analyze the relationship between this long-term bias and visual attention. In our research, we identify a consistent phenomenon in current LVLMs: the model's attention to visual input diminishes as the generated sequence grows, which we hypothesize to be a key factor contributing to observed increasing hallucinations. Based on these insights, we propose Image attention-guided Key-value merging cOllaborative Decoding (IKOD), a collaborative decoding strategy generating more image-focused sequences. This method derives logits from shorter sequences with higher image attention through key-value merging and combines them with those from the original decoding, effectively mitigating attention degradation and suppressing hallucinations while not incurring too much inference cost. Extensive experiments on both hallucination and comprehensive benchmarks demonstrate IKOD's superior effectiveness in mitigating hallucinations and improving comprehensive capacities for LVLMs. Importantly, IKOD requires no additional training or external tools, making it a lightweight and efficient framework applicable to various models.
Authors: Junjie Cao, Kaizhou Li, Xinchun Yu, Hongxiang Li, Xiaoping Zhang
Abstract: With the rapid development of generative technology, current generative models can generate high-fidelity digital content and edit it in a controlled manner. However, there is a risk that malicious individuals might misuse these capabilities for misleading activities. Although existing research has attempted to shield photographic images from being manipulated by generative models, there remains a significant disparity in the protection offered to video content editing. To bridge the gap, we propose a protection method named VideoGuard, which can effectively protect videos from unauthorized malicious editing. This protection is achieved through the subtle introduction of nearly unnoticeable perturbations that interfere with the functioning of the intended generative diffusion models. Due to the redundancy between video frames, and inter-frame attention mechanism in video diffusion models, simply applying image-based protection methods separately to every video frame can not shield video from unauthorized editing. To tackle the above challenge, we adopt joint frame optimization, treating all video frames as an optimization entity. Furthermore, we extract video motion information and fuse it into optimization objectives. Thus, these alterations can effectively force the models to produce outputs that are implausible and inconsistent. We provide a pipeline to optimize this perturbation. Finally, we use both objective metrics and subjective metrics to demonstrate the efficacy of our method, and the results show that the protection performance of VideoGuard is superior to all the baseline methods.
Authors: Hyungjin Kim, Seokho Ahn, Young-Duk Seo
Abstract: Personalized generation in T2I diffusion models aims to naturally incorporate individual user preferences into the generation process with minimal user intervention. However, existing studies primarily rely on prompt-level modeling with large-scale models, often leading to inaccurate personalization due to the limited input token capacity of T2I diffusion models. To address these limitations, we propose DrUM, a novel method that integrates user profiling with a transformer-based adapter to enable personalized generation through condition-level modeling in the latent space. DrUM demonstrates strong performance on large-scale datasets and seamlessly integrates with open-source text encoders, making it compatible with widely used foundation T2I models without requiring additional fine-tuning.
Authors: Dasol Choi Jihwan Lee, Minjae Lee, Minsuk Kahng
Abstract: While prior research on text-to-image generation has predominantly focused on biases in human depictions, we investigate a more subtle yet pervasive phenomenon: demographic bias in generated objects (e.g., cars). We introduce SODA (Stereotyped Object Diagnostic Audit), a novel framework for systematically measuring such biases. Our approach compares visual attributes of objects generated with demographic cues (e.g., "for young people'') to those from neutral prompts, across 2,700 images produced by three state-of-the-art models (GPT Image-1, Imagen 4, and Stable Diffusion) in five object categories. Through a comprehensive analysis, we uncover strong associations between specific demographic groups and visual attributes, such as recurring color patterns prompted by gender or ethnicity cues. These patterns reflect and reinforce not only well-known stereotypes but also more subtle and unintuitive biases. We also observe that some models generate less diverse outputs, which in turn amplifies the visual disparities compared to neutral prompts. Our proposed auditing framework offers a practical approach for testing, revealing how stereotypes still remain embedded in today's generative models. We see this as an essential step toward more systematic and responsible AI development.
Authors: Lianwei Yang, Haokun Lin, Tianchen Zhao, Yichen Wu, Hongyu Zhu, Ruiqi Xie, Zhenan Sun, Yu Wang, Qingyi Gu
Abstract: Diffusion Transformers (DiTs) have achieved impressive performance in text-to-image generation. However, their high computational cost and large parameter sizes pose significant challenges for usage in resource-constrained scenarios. Post-training quantization (PTQ) is a promising solution to reduce memory usage and accelerate inference, but existing PTQ methods suffer from severe performance degradation under extreme low-bit settings. We identify two key obstacles to low-bit post-training quantization for DiT models: (1) model weights follow a Gaussian-like distribution with long tails, causing uniform quantization to poorly allocate intervals and leading to significant errors; (2) two types of activation outliers: (i) Mild Outliers with slightly elevated values, and (ii) Salient Outliers with large magnitudes concentrated in specific channels, which disrupt activation quantization. To address these issues, we propose LRQ-DiT, an efficient and accurate PTQ framework. We introduce Twin-Log Quantization (TLQ), a log-based method that aligns well with the weight distribution and reduces quantization errors. We also propose an Adaptive Rotation Scheme (ARS) that dynamically applies Hadamard or outlier-aware rotations based on activation fluctuation, effectively mitigating the impact of both types of outliers. We evaluate LRQ-DiT on PixArt and FLUX under various bit-width settings, and validate the performance on COCO, MJHQ, and sDCI datasets. LRQ-DiT achieves low-bit quantization of DiT models while preserving image quality, outperforming existing PTQ baselines.
Authors: Yu Zhou, Pelle Thielmann, Ayush Chamoli, Bruno Mirbach, Didier Stricker, Jason Rambach
Abstract: The construction industry represents a major sector in terms of resource consumption. Recycled construction material has high reuse potential, but quality monitoring of the aggregates is typically still performed with manual methods. Vision-based machine learning methods could offer a faster and more efficient solution to this problem, but existing segmentation methods are by design not directly applicable to images with hundreds of small particles. In this paper, we propose ParticleSAM, an adaptation of the segmentation foundation model to images with small and dense objects such as the ones often encountered in construction material particles. Moreover, we create a new dense multi-particle dataset simulated from isolated particle images with the assistance of an automated data generation and labeling pipeline. This dataset serves as a benchmark for visual material quality control automation while our segmentation approach has the potential to be valuable in application areas beyond construction where small-particle segmentation is needed. Our experimental results validate the advantages of our method by comparing to the original SAM method both in quantitative and qualitative experiments.
Authors: Mohammadsadegh Khoshghiaferezaee, Moritz Krauth, Shima Shabani, Michael Breu{\ss}
Abstract: Sparse dictionary learning (SDL) is a fundamental technique that is useful for many image processing tasks. As an example we consider here image recovery, where SDL can be cast as a nonsmooth optimization problem. For this kind of problems, iterative shrinkage methods represent a powerful class of algorithms that are subject of ongoing research. Sparsity is an important property of the learned solutions, as exactly the sparsity enables efficient further processing or storage. The sparsity implies that a recovered image is determined as a combination of a number of dictionary elements that is as low as possible. Therefore, the question arises, to which degree sparsity should be enforced in SDL in order to not compromise recovery quality. In this paper we focus on the sparsity of solutions that can be obtained using a variety of optimization methods. It turns out that there are different sparsity regimes depending on the method in use. Furthermore, we illustrate that high sparsity does in general not compromise recovery quality, even if the recovered image is quite different from the learning database.
Authors: Shreyank N Gowda, Xiaobo Jin, Christian Wagner
Abstract: In cross-modal retrieval tasks, such as image-to-report and report-to-image retrieval, accurately aligning medical images with relevant text reports is essential but challenging due to the inherent ambiguity and variability in medical data. Existing models often struggle to capture the nuanced, multi-level semantic relationships in radiology data, leading to unreliable retrieval results. To address these issues, we propose the Prototype-Enhanced Confidence Modeling (PECM) framework, which introduces multi-level prototypes for each modality to better capture semantic variability and enhance retrieval robustness. PECM employs a dual-stream confidence estimation that leverages prototype similarity distributions and an adaptive weighting mechanism to control the impact of high-uncertainty data on retrieval rankings. Applied to radiology image-report datasets, our method achieves significant improvements in retrieval precision and consistency, effectively handling data ambiguity and advancing reliability in complex clinical scenarios. We report results on multiple different datasets and tasks including fully supervised and zero-shot retrieval obtaining performance gains of up to 10.17%, establishing in new state-of-the-art.
Authors: Deqiang Yin, Junyi Guo, Huanda Lu, Fangyu Wu, Dongming Lu
Abstract: Instruction-based garment editing enables precise image modifications via natural language, with broad applications in fashion design and customization. Unlike general editing tasks, it requires understanding garment-specific semantics and attribute dependencies. However, progress is limited by the scarcity of high-quality instruction-image pairs, as manual annotation is costly and hard to scale. While MLLMs have shown promise in automated data synthesis, their application to garment editing is constrained by imprecise instruction modeling and a lack of fashion-specific supervisory signals. To address these challenges, we present an automated pipeline for constructing a garment editing dataset. We first define six editing instruction categories aligned with real-world fashion workflows to guide the generation of balanced and diverse instruction-image triplets. Second, we introduce Fashion Edit Score, a semantic-aware evaluation metric that captures semantic dependencies between garment attributes and provides reliable supervision during construction. Using this pipeline, we construct a total of 52,257 candidate triplets and retain 20,596 high-quality triplets to build EditGarment, the first instruction-based dataset tailored to standalone garment editing. The project page is https://yindq99.github.io/EditGarment-project/.
Authors: Yazhou Zhu, Haofeng Zhang
Abstract: Cross-domain Few-shot Medical Image Segmentation (CD-FSMIS) is a potential solution for segmenting medical images with limited annotation using knowledge from other domains. The significant performance of current CD-FSMIS models relies on the heavily training procedure over other source medical domains, which degrades the universality and ease of model deployment. With the development of large visual models of natural images, we propose a training-free CD-FSMIS model that introduces the Multi-center Adaptive Uncertainty-aware Prompting (MAUP) strategy for adapting the foundation model Segment Anything Model (SAM), which is trained with natural images, into the CD-FSMIS task. To be specific, MAUP consists of three key innovations: (1) K-means clustering based multi-center prompts generation for comprehensive spatial coverage, (2) uncertainty-aware prompts selection that focuses on the challenging regions, and (3) adaptive prompt optimization that can dynamically adjust according to the target region complexity. With the pre-trained DINOv2 feature encoder, MAUP achieves precise segmentation results across three medical datasets without any additional training compared with several conventional CD-FSMIS models and training-free FSMIS model. The source code is available at: https://github.com/YazhouZhu19/MAUP.
Authors: Shiben Liu, Mingyue Xu, Huijie Fan, Qiang Wang, Yandong Tang, Zhi Han
Abstract: Lifelong person re-identification (LReID) encounters a key challenge: balancing the preservation of old knowledge with adaptation to new information. Existing LReID methods typically employ knowledge distillation to enforce representation alignment. However, these approaches ignore two crucial aspects: specific distribution awareness and cross-domain unified knowledge learning, both of which are essential for addressing this challenge. To overcome these limitations, we propose a novel distribution-aware knowledge unification and association (DKUA) framework where domain-style modeling is performed for each instance to propagate domain-specific representations, enhancing anti-forgetting and generalization capacity. Specifically, we design a distribution-aware model to transfer instance-level representations of the current domain into the domain-specific representations with the different domain styles, preserving learned knowledge without storing old samples. Next, we propose adaptive knowledge consolidation (AKC) to dynamically generate the unified representation as a cross-domain representation center. To further mitigate forgetting, we develop a unified knowledge association (UKA) mechanism, which explores the unified representation as a bridge to explicitly model inter-domain associations, reducing inter-domain gaps. Finally, distribution-based knowledge transfer (DKT) is proposed to prevent the current domain distribution from deviating from the cross-domain distribution center, improving adaptation capacity. Experimental results show our DKUA outperforms the existing methods by 7.6%/5.3% average mAP/R@1 improvement on anti-forgetting and generalization capacity, respectively. Our code will be publicly released.
Authors: Stefan Brandst\"atter, Maximilian K\"oller, Philipp Seeb\"ock, Alissa Blessing, Felicitas Oberndorfer, Svitlana Pochepnia, Helmut Prosch, Georg Langs
Abstract: In histopathology, tissue samples are often larger than a standard microscope slide, making stitching of multiple fragments necessary to process entire structures such as tumors. Automated stitching is a prerequisite for scaling analysis, but is challenging due to possible tissue loss during preparation, inhomogeneous morphological distortion, staining inconsistencies, missing regions due to misalignment on the slide, or frayed tissue edges. This limits state-of-the-art stitching methods using boundary shape matching algorithms to reconstruct artificial whole mount slides (WMS). Here, we introduce SemanticStitcher using latent feature representations derived from a visual histopathology foundation model to identify neighboring areas in different fragments. Robust pose estimation based on a large number of semantic matching candidates derives a mosaic of multiple fragments to form the WMS. Experiments on three different histopathology datasets demonstrate that SemanticStitcher yields robust WMS mosaicing and consistently outperforms the state of the art in correct boundary matches.
Authors: Kaishen Yuan, Yuting Zhang, Shang Gao, Yijie Zhu, Wenshuo Chen, Yutao Yue
Abstract: Emotional Image Content Generation (EICG) aims to generate semantically clear and emotionally faithful images based on given emotion categories, with broad application prospects. While recent text-to-image diffusion models excel at generating concrete concepts, they struggle with the complexity of abstract emotions. There have also emerged methods specifically designed for EICG, but they excessively rely on word-level attribute labels for guidance, which suffer from semantic incoherence, ambiguity, and limited scalability. To address these challenges, we propose CoEmoGen, a novel pipeline notable for its semantic coherence and high scalability. Specifically, leveraging multimodal large language models (MLLMs), we construct high-quality captions focused on emotion-triggering content for context-rich semantic guidance. Furthermore, inspired by psychological insights, we design a Hierarchical Low-Rank Adaptation (HiLoRA) module to cohesively model both polarity-shared low-level features and emotion-specific high-level semantics. Extensive experiments demonstrate CoEmoGen's superiority in emotional faithfulness and semantic coherence from quantitative, qualitative, and user study perspectives. To intuitively showcase scalability, we curate EmoArt, a large-scale dataset of emotionally evocative artistic images, providing endless inspiration for emotion-driven artistic creation. The dataset and code are available at https://github.com/yuankaishen2001/CoEmoGen.
Authors: Inamullah, Imran Razzak, Shoaib Jameel
Abstract: Retinal microvascular imaging is increasingly recognised as a non invasive method for evaluating systemic vascular and metabolic health. However, the association between lipidomics and retinal vasculature remains inadequate. This study investigates the relationships between serum lipid subclasses, free fatty acids (FA), diacylglycerols (DAG), triacylglycerols (TAG), and cholesteryl esters (CE), and retinal microvascular characteristics in a large population-based cohort. Using Spearman correlation analysis, we examined the interconnection between lipid subclasses and ten retinal microvascular traits, applying the Benjamini-Hochberg false discovery rate (BH-FDR) to adjust for statistical significance. Results indicated that FA were linked to retinal vessel twistiness, while CE correlated with the average widths of arteries and veins. Conversely, DAG and TAG showed negative correlations with the width and complexity of arterioles and venules. These findings suggest that retinal vascular architecture reflects distinct circulating lipid profiles, supporting its role as a non-invasive marker of systemic metabolic health. This study is the first to integrate deep learning (DL)derived retinal traits with lipidomic subclasses in a healthy cohort, thereby providing insights into microvascular structural changes independent of disease status or treatment effects.
Authors: Long Qian, Bingke Zhu, Yingying Chen, Ming Tang, Jinqiao Wang
Abstract: Despite substantial progress in anomaly synthesis methods, existing diffusion-based and coarse inpainting pipelines commonly suffer from structural deficiencies such as micro-structural discontinuities, limited semantic controllability, and inefficient generation. To overcome these limitations, we introduce ARAS, a language-conditioned, auto-regressive anomaly synthesis approach that precisely injects local, text-specified defects into normal images via token-anchored latent editing. Leveraging a hard-gated auto-regressive operator and a training-free, context-preserving masked sampling kernel, ARAS significantly enhances defect realism, preserves fine-grained material textures, and provides continuous semantic control over synthesized anomalies. Integrated within our Quality-Aware Re-weighted Anomaly Detection (QARAD) framework, we further propose a dynamic weighting strategy that emphasizes high-quality synthetic samples by computing an image-text similarity score with a dual-encoder model. Extensive experiments across three benchmark datasets-MVTec AD, VisA, and BTAD, demonstrate that our QARAD outperforms SOTA methods in both image- and pixel-level anomaly detection tasks, achieving improved accuracy, robustness, and a 5 times synthesis speedup compared to diffusion-based alternatives. Our complete code and synthesized dataset will be publicly available.
Authors: Dmitrii Korzh, Dmitrii Tarasov, Artyom Iudin, Elvir Karimov, Matvey Skripkin, Nikita Kuzmin, Andrey Kuznetsov, Oleg Y. Rogov, Ivan Oseledets
Abstract: Conversion of spoken mathematical expressions is a challenging task that involves transcribing speech into a strictly structured symbolic representation while addressing the ambiguity inherent in the pronunciation of equations. Although significant progress has been achieved in automatic speech recognition (ASR) and language models (LM), the problem of converting spoken mathematics into LaTeX remains underexplored. This task directly applies to educational and research domains, such as lecture transcription or note creation. Based on ASR post-correction, prior work requires 2 transcriptions, focuses only on isolated equations, has a limited test set, and provides neither training data nor multilingual coverage. To address these issues, we present the first fully open-source large-scale dataset, comprising over 66,000 human-annotated audio samples of mathematical equations and sentences in both English and Russian, drawn from diverse scientific domains. In addition to the ASR post-correction models and few-shot prompting, we apply audio language models, demonstrating comparable character error rate (CER) results on the MathSpeech benchmark (28% vs. 30%) for the equations conversion. In contrast, on the proposed S2L-equations benchmark, our models outperform the MathSpeech model by a substantial margin of more than 40 percentage points, even after accounting for LaTeX formatting artifacts (27% vs. 64%). We establish the first benchmark for mathematical sentence recognition (S2L-sentences) and achieve an equation CER of 40%. This work lays the groundwork for future advances in multimodal AI, with a particular focus on mathematical content recognition.
Authors: Stephanie Wohlfahrt, Christoph Praschl, Horst Leitner, Wolfram Jantsch, Julia Konic, Silvio Schueler, Andreas St\"ockl, David C. Schedl
Abstract: We use unmanned aerial drones to estimate wildlife density in southeastern Austria and compare these estimates to camera trap data. Traditional methods like capture-recapture, distance sampling, or camera traps are well-established but labour-intensive or spatially constrained. Using thermal (IR) and RGB imagery, drones enable efficient, non-intrusive animal counting. Our surveys were conducted during the leafless period on single days in October and November 2024 in three areas of a sub-Illyrian hill and terrace landscape. Flight transects were based on predefined launch points using a 350 m grid and an algorithm that defined the direction of systematically randomized transects. This setup allowed surveying large areas in one day using multiple drones, minimizing double counts. Flight altitude was set at 60 m to avoid disturbing roe deer (Capreolus capreolus) while ensuring detection. Animals were manually annotated in the recorded imagery and extrapolated to densities per square kilometer. We applied three extrapolation methods with increasing complexity: naive area-based extrapolation, bootstrapping, and zero-inflated negative binomial modelling. For comparison, a Random Encounter Model (REM) estimate was calculated using camera trap data from the flight period. The drone-based methods yielded similar results, generally showing higher densities than REM, except in one area in October. We hypothesize that drone-based density reflects daytime activity in open and forested areas, while REM estimates average activity over longer periods within forested zones. Although both approaches estimate density, they offer different perspectives on wildlife presence. Our results show that drones offer a promising, scalable method for wildlife density estimation.
Authors: Muzhaffar Hazman, Susan McKeever, Josephine Griffith
Abstract: Internet memes, now a staple of digital communication, play a pivotal role in how users engage within online communities and allow researchers to gain insight into contemporary digital culture. These engaging user-generated content are characterised by their reuse of visual elements also found in other memes. Matching instances of memes via these shared visual elements, called Meme Matching, is the basis of a wealth of meme analysis approaches. However, most existing methods assume that every meme consists of a shared visual background, called a Template, with some overlaid text, thereby limiting meme matching to comparing the background image alone. Current approaches exclude the many memes that are not template-based and limit the effectiveness of automated meme analysis and would not be effective at linking memes to contemporary web-based meme dictionaries. In this work, we introduce a broader formulation of meme matching that extends beyond template matching. We show that conventional similarity measures, including a novel segment-wise computation of the similarity measures, excel at matching template-based memes but fall short when applied to non-template-based meme formats. However, the segment-wise approach was found to consistently outperform the whole-image measures on matching non-template-based memes. Finally, we explore a prompting-based approach using a pretrained Multimodal Large Language Model for meme matching. Our results highlight that accurately matching memes via shared visual elements, not just background templates, remains an open challenge that requires more sophisticated matching techniques.
Authors: Annemarie McCarthy
Abstract: Historical maps offer a valuable lens through which to study past landscapes and settlement patterns. While prior research has leveraged machine learning based techniques to extract building footprints from historical maps, such approaches have largely focused on urban areas and tend to be computationally intensive. This presents a challenge for research questions requiring analysis across extensive rural regions, such as verifying historical census data or locating abandoned settlements. In this paper, this limitation is addressed by proposing a scalable and efficient pipeline tailored to rural maps with sparse building distributions. The method described employs a hierarchical machine learning based approach: convolutional neural network (CNN) classifiers are first used to progressively filter out map sections unlikely to contain buildings, significantly reducing the area requiring detailed analysis. The remaining high probability sections are then processed using CNN segmentation algorithms to extract building features. The pipeline is validated using test sections from the Ordnance Survey Ireland historical 25 inch map series and 6 inch map series, demonstrating both high performance and improved efficiency compared to conventional segmentation-only approaches. Application of the technique to both map series, covering the same geographic region, highlights its potential for historical and archaeological discovery. Notably, the pipeline identified a settlement of approximately 22 buildings in Tully, Co. Galway, present in the 6 inch map, produced in 1839, but absent from the 25 inch map, produced in 1899, suggesting it may have been abandoned during the Great Famine period.
Authors: Xinyu Xiong, Zihuang Wu, Lei Zhang, Lei Lu, Ming Li, Guanbin Li
Abstract: Recent studies have highlighted the potential of adapting the Segment Anything Model (SAM) for various downstream tasks. However, constructing a more powerful and generalizable encoder to further enhance performance remains an open challenge. In this work, we propose SAM2-UNeXT, an advanced framework that builds upon the core principles of SAM2-UNet while extending the representational capacity of SAM2 through the integration of an auxiliary DINOv2 encoder. By incorporating a dual-resolution strategy and a dense glue layer, our approach enables more accurate segmentation with a simple architecture, relaxing the need for complex decoder designs. Extensive experiments conducted on four benchmarks, including dichotomous image segmentation, camouflaged object detection, marine animal segmentation, and remote sensing saliency detection, demonstrate the superior performance of our proposed method. The code is available at https://github.com/WZH0120/SAM2-UNeXT.
Authors: Jonas Leo Mueller, Lukas Engel, Eva Dorschky, Daniel Krauss, Ingrid Ullmann, Martin Vossiek, Bjoern M. Eskofier
Abstract: Radar-based human pose estimation (HPE) provides a privacy-preserving, illumination-invariant sensing modality but is challenged by noisy, multipath-affected measurements. We introduce RadProPoser, a probabilistic encoder-decoder architecture that processes complex-valued radar tensors from a compact 3-transmitter, 4-receiver MIMO radar. By incorporating variational inference into keypoint regression, RadProPoser jointly predicts 26 three-dimensional joint locations alongside heteroscedastic aleatoric uncertainties and can be recalibrated to predict total uncertainty. We explore different probabilistic formulations using both Gaussian and Laplace distributions for latent priors and likelihoods. On our newly released dataset with optical motion-capture ground truth, RadProPoser achieves an overall mean per-joint position error (MPJPE) of 6.425 cm, with 5.678 cm at the 45 degree aspect angle. The learned uncertainties exhibit strong alignment with actual pose errors and can be calibrated to produce reliable prediction intervals, with our best configuration achieving an expected calibration error of 0.021. As an additional demonstration, sampling from these latent distributions enables effective data augmentation for downstream activity classification, resulting in an F1 score of 0.870. To our knowledge, this is the first end-to-end radar tensor-based HPE system to explicitly model and quantify per-joint uncertainty from raw radar tensor data, establishing a foundation for explainable and reliable human motion analysis in radar applications.
Authors: Wuyang Li, Wentao Pan, Xiaoyuan Liu, Zhendong Luo, Chenxin Li, Hengyu Liu, Din Ping Tsai, Mu Ku Chen, Yixuan Yuan
Abstract: Miniaturized endoscopy has advanced accurate visual perception within the human body. Prevailing research remains limited to conventional cameras employing convex lenses, where the physical constraints with millimetre-scale thickness impose serious impediments on the micro-level clinical. Recently, with the emergence of meta-optics, ultra-micro imaging based on metalenses (micron-scale) has garnered great attention, serving as a promising solution. However, due to the physical difference of metalens, there is a large gap in data acquisition and algorithm research. In light of this, we aim to bridge this unexplored gap, advancing the novel metalens endoscopy. First, we establish datasets for metalens endoscopy and conduct preliminary optical simulation, identifying two derived optical issues that physically adhere to strong optical priors. Second, we propose MetaScope, a novel optics-driven neural network tailored for metalens endoscopy driven by physical optics. MetaScope comprises two novel designs: Optics-informed Intensity Adjustment (OIA), rectifying intensity decay by learning optical embeddings, and Optics-informed Chromatic Correction (OCC), mitigating chromatic aberration by learning spatial deformations informed by learned Point Spread Function (PSF) distributions. To enhance joint learning, we further deploy a gradient-guided distillation to transfer knowledge from the foundational model adaptively. Extensive experiments demonstrate that MetaScope not only outperforms state-of-the-art methods in both metalens segmentation and restoration but also achieves impressive generalized ability in real biomedical scenes.
Authors: Md Abrar Jahin, Shahriar Soudeep, M. F. Mridha, Nafiz Fahad, Md. Jakir Hossen
Abstract: Recent advancements in object detection rely on modular architectures with multi-scale fusion and attention mechanisms. However, static fusion heuristics and class-agnostic attention limit performance in dynamic scenes with occlusions, clutter, and class imbalance. We introduce Dynamic Class-Aware Fusion Network (DyCAF-Net) that addresses these challenges through three innovations: (1) an input-conditioned equilibrium-based neck that iteratively refines multi-scale features via implicit fixed-point modeling, (2) a dual dynamic attention mechanism that adaptively recalibrates channel and spatial responses using input- and class-dependent cues, and (3) class-aware feature adaptation that modulates features to prioritize discriminative regions for rare classes. Through comprehensive ablation studies with YOLOv8 and related architectures, alongside benchmarking against nine state-of-the-art baselines, DyCAF-Net achieves significant improvements in precision, mAP@50, and mAP@50-95 across 13 diverse benchmarks, including occlusion-heavy and long-tailed datasets. The framework maintains computational efficiency ($\sim$11.1M parameters) and competitive inference speeds, while its adaptability to scale variance, semantic overlaps, and class imbalance positions it as a robust solution for real-world detection tasks in medical imaging, surveillance, and autonomous systems.
Authors: Saleh Sakib Ahmed, Sara Nowreen, M. Sohel Rahman
Abstract: Cloud cover and nighttime conditions remain significant limitations in satellite-based remote sensing, often restricting the availability and usability of multi-spectral imagery. In contrast, Sentinel-1 radar images are unaffected by cloud cover and can provide consistent data regardless of weather or lighting conditions. To address the challenges of limited satellite imagery, we propose CloudBreaker, a novel framework that generates high-quality multi-spectral Sentinel-2 signals from Sentinel-1 data. This includes the reconstruction of optical (RGB) images as well as critical vegetation and water indices such as NDVI and NDWI.We employed a novel multi-stage training approach based on conditional latent flow matching and, to the best of our knowledge, are the first to integrate cosine scheduling with flow matching. CloudBreaker demonstrates strong performance, achieving a Frechet Inception Distance (FID) score of 0.7432, indicating high fidelity and realism in the generated optical imagery. The model also achieved Structural Similarity Index Measure (SSIM) of 0.6156 for NDWI and 0.6874 for NDVI, indicating a high degree of structural similarity. This establishes CloudBreaker as a promising solution for a wide range of remote sensing applications where multi-spectral data is typically unavailable or unreliable
Authors: Rodrigo Verschae, Ignacio Bugueno-Cordova
Abstract: Event-based cameras are bio-inspired vision sensors that asynchronously capture per-pixel intensity changes with microsecond latency, high temporal resolution, and high dynamic range, providing valuable information about the spatio-temporal dynamics of the scene. In the present work, we propose evTransFER, a transfer learning-based framework and architecture for face expression recognition using event-based cameras. The main contribution is a feature extractor designed to encode the spatio-temporal dynamics of faces, built by training an adversarial generative method on a different problem (facial reconstruction) and then transferring the trained encoder weights to the face expression recognition system. We show that this proposed transfer learning method greatly improves the ability to recognize facial expressions compared to training a network from scratch. In addition, we propose an architecture that incorporates an LSTM to capture longer-term facial expression dynamics, and we introduce a new event-based representation, referred to as TIE, both of which further improve the results. We evaluate the proposed framework on the event-based facial expression database e-CK+ and compare it to state-of-the-art methods. The results show that the proposed framework evTransFER achieves a 93.6\% recognition rate on the e-CK+ database, significantly improving the accuracy (25.9\% points or more) when compared to state-of-the-art performance for similar problems.
Authors: Nassim Ali Ousalah, Peyman Rostami, Anis Kacem, Enjie Ghorbel, Emmanuel Koumandakis, Djamila Aouada
Abstract: We introduce FPG-NAS, a FLOPs-aware Gated Differentiable Neural Architecture Search framework for efficient 6DoF object pose estimation. Estimating 3D rotation and translation from a single image has been widely investigated yet remains computationally demanding, limiting applicability in resource-constrained scenarios. FPG-NAS addresses this by proposing a specialized differentiable NAS approach for 6DoF pose estimation, featuring a task-specific search space and a differentiable gating mechanism that enables discrete multi-candidate operator selection, thus improving architectural diversity. Additionally, a FLOPs regularization term ensures a balanced trade-off between accuracy and efficiency. The framework explores a vast search space of approximately 10\textsuperscript{92} possible architectures. Experiments on the LINEMOD and SPEED+ datasets demonstrate that FPG-NAS-derived models outperform previous methods under strict FLOPs constraints. To the best of our knowledge, FPG-NAS is the first differentiable NAS framework specifically designed for 6DoF object pose estimation.
Authors: Daniel DeAlcala, Aythami Morales, Julian Fierrez, Ruben Tolosana
Abstract: We present Attention Zoom, a modular and model-agnostic spatial attention mechanism designed to improve feature extraction in convolutional neural networks (CNNs). Unlike traditional attention approaches that require architecture-specific integration, our method introduces a standalone layer that spatially emphasizes high-importance regions in the input. We evaluated Attention Zoom on multiple CNN backbones using CIFAR-100 and TinyImageNet, showing consistent improvements in Top-1 and Top-5 classification accuracy. Visual analyses using Grad-CAM and spatial warping reveal that our method encourages fine-grained and diverse attention patterns. Our results confirm the effectiveness and generality of the proposed layer for improving CCNs with minimal architectural overhead.
Authors: Xiangyu Sun, Haoyi jiang, Liu Liu, Seungtae Nam, Gyeongjin Kang, Xinjie wang, Wei Sui, Zhizhong Su, Wenyu Liu, Xinggang Wang, Eunbyung Park
Abstract: Reconstructing and semantically interpreting 3D scenes from sparse 2D views remains a fundamental challenge in computer vision. Conventional methods often decouple semantic understanding from reconstruction or necessitate costly per-scene optimization, thereby restricting their scalability and generalizability. In this paper, we introduce Uni3R, a novel feed-forward framework that jointly reconstructs a unified 3D scene representation enriched with open-vocabulary semantics, directly from unposed multi-view images. Our approach leverages a Cross-View Transformer to robustly integrate information across arbitrary multi-view inputs, which then regresses a set of 3D Gaussian primitives endowed with semantic feature fields. This unified representation facilitates high-fidelity novel view synthesis, open-vocabulary 3D semantic segmentation, and depth prediction, all within a single, feed-forward pass. Extensive experiments demonstrate that Uni3R establishes a new state-of-the-art across multiple benchmarks, including 25.07 PSNR on RE10K and 55.84 mIoU on ScanNet. Our work signifies a novel paradigm towards generalizable, unified 3D scene reconstruction and understanding. The code is available at https://github.com/HorizonRobotics/Uni3R.
Authors: Katherine Liu, Sergey Zakharov, Dian Chen, Takuya Ikeda, Greg Shakhnarovich, Adrien Gaidon, Rares Ambrus
Abstract: We would like to estimate the pose and full shape of an object from a single observation, without assuming known 3D model or category. In this work, we propose OmniShape, the first method of its kind to enable probabilistic pose and shape estimation. OmniShape is based on the key insight that shape completion can be decoupled into two multi-modal distributions: one capturing how measurements project into a normalized object reference frame defined by the dataset and the other modelling a prior over object geometries represented as triplanar neural fields. By training separate conditional diffusion models for these two distributions, we enable sampling multiple hypotheses from the joint pose and shape distribution. OmniShape demonstrates compelling performance on challenging real world datasets. Project website: https://tri-ml.github.io/omnishape
Authors: Youquan Liu, Lingdong Kong, Weidong Yang, Ao Liang, Jianxiong Gao, Yang Wu, Xiang Xu, Xin Li, Linfeng Li, Runnan Chen, Ben Fei
Abstract: Realistic and controllable panoramic LiDAR data generation is critical for scalable 3D perception in autonomous driving and robotics. Existing methods either perform unconditional generation with poor controllability or adopt text-guided synthesis, which lacks fine-grained spatial control. Leveraging a monocular RGB image as a spatial control signal offers a scalable and low-cost alternative, which remains an open problem. However, it faces three core challenges: (i) semantic and depth cues from RGB are vary spatially, complicating reliable conditioning generation; (ii) modality gaps between RGB appearance and LiDAR geometry amplify alignment errors under noisy diffusion; and (iii) maintaining structural coherence between monocular RGB and panoramic LiDAR is challenging, particularly in non-overlap regions between images and LiDAR. To address these challenges, we propose Veila, a novel conditional diffusion framework that integrates: a Confidence-Aware Conditioning Mechanism (CACM) that strengthens RGB conditioning by adaptively balancing semantic and depth cues according to their local reliability; a Geometric Cross-Modal Alignment (GCMA) for robust RGB-LiDAR alignment under noisy diffusion; and a Panoramic Feature Coherence (PFC) for enforcing global structural consistency across monocular RGB and panoramic LiDAR. Additionally, we introduce two metrics, Cross-Modal Semantic Consistency and Cross-Modal Depth Consistency, to evaluate alignment quality across modalities. Experiments on nuScenes, SemanticKITTI, and our proposed KITTI-Weather benchmark demonstrate that Veila achieves state-of-the-art generation fidelity and cross-modal consistency, while enabling generative data augmentation that improves downstream LiDAR semantic segmentation.
Authors: Youquan Liu, Lingdong Kong, Weidong Yang, Xin Li, Ao Liang, Runnan Chen, Ben Fei, Tongliang Liu
Abstract: Controllable generation of realistic LiDAR scenes is crucial for applications such as autonomous driving and robotics. While recent diffusion-based models achieve high-fidelity LiDAR generation, they lack explicit control over foreground objects and spatial relationships, limiting their usefulness for scenario simulation and safety validation. To address these limitations, we propose Large-scale Layout-guided LiDAR generation model ("La La LiDAR"), a novel layout-guided generative framework that introduces semantic-enhanced scene graph diffusion with relation-aware contextual conditioning for structured LiDAR layout generation, followed by foreground-aware control injection for complete scene generation. This enables customizable control over object placement while ensuring spatial and semantic consistency. To support our structured LiDAR generation, we introduce Waymo-SG and nuScenes-SG, two large-scale LiDAR scene graph datasets, along with new evaluation metrics for layout synthesis. Extensive experiments demonstrate that La La LiDAR achieves state-of-the-art performance in both LiDAR generation and downstream perception tasks, establishing a new benchmark for controllable 3D scene generation.
Authors: Ao Liang, Youquan Liu, Yu Yang, Dongyue Lu, Linfeng Li, Lingdong Kong, Huaici Zhao, Wei Tsang Ooi
Abstract: Generative world models have become essential data engines for autonomous driving, yet most existing efforts focus on videos or occupancy grids, overlooking the unique LiDAR properties. Extending LiDAR generation to dynamic 4D world modeling presents challenges in controllability, temporal coherence, and evaluation standardization. To this end, we present LiDARCrafter, a unified framework for 4D LiDAR generation and editing. Given free-form natural language inputs, we parse instructions into ego-centric scene graphs, which condition a tri-branch diffusion network to generate object structures, motion trajectories, and geometry. These structured conditions enable diverse and fine-grained scene editing. Additionally, an autoregressive module generates temporally coherent 4D LiDAR sequences with smooth transitions. To support standardized evaluation, we establish a comprehensive benchmark with diverse metrics spanning scene-, object-, and sequence-level aspects. Experiments on the nuScenes dataset using this benchmark demonstrate that LiDARCrafter achieves state-of-the-art performance in fidelity, controllability, and temporal consistency across all levels, paving the way for data augmentation and simulation. The code and benchmark are released to the community.
Authors: Jianxiong Gao, Zhaoxi Chen, Xian Liu, Jianfeng Feng, Chenyang Si, Yanwei Fu, Yu Qiao, Ziwei Liu
Abstract: Controllable ultra-long video generation is a fundamental yet challenging task. Although existing methods are effective for short clips, they struggle to scale due to issues such as temporal inconsistency and visual degradation. In this paper, we initially investigate and identify three key factors: separate noise initialization, independent control signal normalization, and the limitations of single-modality guidance. To address these issues, we propose LongVie, an end-to-end autoregressive framework for controllable long video generation. LongVie introduces two core designs to ensure temporal consistency: 1) a unified noise initialization strategy that maintains consistent generation across clips, and 2) global control signal normalization that enforces alignment in the control space throughout the entire video. To mitigate visual degradation, LongVie employs 3) a multi-modal control framework that integrates both dense (e.g., depth maps) and sparse (e.g., keypoints) control signals, complemented by 4) a degradation-aware training strategy that adaptively balances modality contributions over time to preserve visual quality. We also introduce LongVGenBench, a comprehensive benchmark consisting of 100 high-resolution videos spanning diverse real-world and synthetic environments, each lasting over one minute. Extensive experiments show that LongVie achieves state-of-the-art performance in long-range controllability, consistency, and quality.
Authors: Pulkit Kumar, Shuaiyi Huang, Matthew Walmer, Sai Saketh Rambhatla, Abhinav Shrivastava
Abstract: Video understanding requires effective modeling of both motion and appearance information, particularly for few-shot action recognition. While recent advances in point tracking have been shown to improve few-shot action recognition, two fundamental challenges persist: selecting informative points to track and effectively modeling their motion patterns. We present Trokens, a novel approach that transforms trajectory points into semantic-aware relational tokens for action recognition. First, we introduce a semantic-aware sampling strategy to adaptively distribute tracking points based on object scale and semantic relevance. Second, we develop a motion modeling framework that captures both intra-trajectory dynamics through the Histogram of Oriented Displacements (HoD) and inter-trajectory relationships to model complex action patterns. Our approach effectively combines these trajectory tokens with semantic features to enhance appearance features with motion information, achieving state-of-the-art performance across six diverse few-shot action recognition benchmarks: Something-Something-V2 (both full and small splits), Kinetics, UCF101, HMDB51, and FineGym. For project page see https://trokens-iccv25.github.io
Authors: Petteri Teikari, Mike Jarrell, Maryam Azh, Harri Pesola
Abstract: The Uniform Appraisal Dataset (UAD) 3.6's mandatory 2026 implementation transforms residential property valuation from narrative reporting to structured, machine-readable formats. This paper provides the first comprehensive analysis of this regulatory shift alongside concurrent AI advances in computer vision, natural language processing, and autonomous systems. We develop a three-layer framework for AI-augmented valuation addressing technical implementation and institutional trust requirements. Our analysis reveals how regulatory standardization converging with AI capabilities enables fundamental market restructuring with profound implications for professional practice, efficiency, and systemic risk. We make four key contributions: (1) documenting institutional failures including inter-appraiser variability and systematic biases undermining valuation reliability; (2) developing an architectural framework spanning physical data acquisition, semantic understanding, and cognitive reasoning that integrates emerging technologies while maintaining professional oversight; (3) addressing trust requirements for high-stakes financial applications including regulatory compliance, algorithmic fairness, and uncertainty quantification; (4) proposing evaluation methodologies beyond generic AI benchmarks toward domain-specific protocols. Our findings indicate successful transformation requires not merely technological sophistication but careful human-AI collaboration, creating systems that augment rather than replace professional expertise while addressing historical biases and information asymmetries in real estate markets.
Authors: Pengwei Sun, Wei Peng, Lun Yu Li, Yixin Wang, Kilian M. Pohl
Abstract: Counterfactual generation offers a principled framework for simulating hypothetical changes in medical imaging, with potential applications in understanding disease mechanisms and generating physiologically plausible data. However, generating realistic structural 3D brain MRIs that respect anatomical and causal constraints remains challenging due to data scarcity, structural complexity, and the lack of standardized evaluation protocols. In this work, we convert six generative models into 3D counterfactual approaches by incorporating an anatomy-guided framework based on a causal graph, in which regional brain volumes serve as direct conditioning inputs. Each model is evaluated with respect to composition, reversibility, realism, effectiveness and minimality on T1-weighted brain MRIs (T1w MRIs) from the Alzheimer's Disease Neuroimaging Initiative (ADNI). In addition, we test the generalizability of each model with respect to T1w MRIs of the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA). Our results indicate that anatomically grounded conditioning successfully modifies the targeted anatomical regions; however, it exhibits limitations in preserving non-targeted structures. Beyond laying the groundwork for more interpretable and clinically relevant generative modeling of brain MRIs, this benchmark highlights the need for novel architectures that more accurately capture anatomical interdependencies.
Authors: Farzad Beizaee, Sina Hajimiri, Ismail Ben Ayed, Gregory Lodygensky, Christian Desrosiers, Jose Dolz
Abstract: Unsupervised anomaly detection (UAD) in brain imaging is crucial for identifying pathologies without the need for labeled data. However, accurately localizing anomalies remains challenging due to the intricate structure of brain anatomy and the scarcity of abnormal examples. In this work, we introduce REFLECT, a novel framework that leverages rectified flows to establish a direct, linear trajectory for correcting abnormal MR images toward a normal distribution. By learning a straight, one-step correction transport map, our method efficiently corrects brain anomalies and can precisely localize anomalies by detecting discrepancies between anomalous input and corrected counterpart. In contrast to the diffusion-based UAD models, which require iterative stochastic sampling, rectified flows provide a direct transport map, enabling single-step inference. Extensive experiments on popular UAD brain segmentation benchmarks demonstrate that REFLECT significantly outperforms state-of-the-art unsupervised anomaly detection methods. The code is available at https://github.com/farzad-bz/REFLECT.
Authors: Puzhen Wu, Mingquan Lin, Qingyu Chen, Emily Y. Chew, Zhiyong Lu, Yifan Peng, Hexin Dong
Abstract: Age-related macular degeneration (AMD) is a leading cause of irreversible vision loss, making effective prognosis crucial for timely intervention. In this work, we propose AMD-Mamba, a novel multi-modal framework for AMD prognosis, and further develop a new AMD biomarker. This framework integrates color fundus images with genetic variants and socio-demographic variables. At its core, AMD-Mamba introduces an innovative metric learning strategy that leverages AMD severity scale score as prior knowledge. This strategy allows the model to learn richer feature representations by aligning learned features with clinical phenotypes, thereby improving the capability of conventional prognosis methods in capturing disease progression patterns. In addition, unlike existing models that use traditional CNN backbones and focus primarily on local information, such as the presence of drusen, AMD-Mamba applies Vision Mamba and simultaneously fuses local and long-range global information, such as vascular changes. Furthermore, we enhance prediction performance through multi-scale fusion, combining image information with clinical variables at different resolutions. We evaluate AMD-Mamba on the AREDS dataset, which includes 45,818 color fundus photographs, 52 genetic variants, and 3 socio-demographic variables from 2,741 subjects. Our experimental results demonstrate that our proposed biomarker is one of the most significant biomarkers for the progression of AMD. Notably, combining this biomarker with other existing variables yields promising improvements in detecting high-risk AMD patients at early stages. These findings highlight the potential of our multi-modal framework to facilitate more precise and proactive management of AMD.
Authors: Brennen A. Hill, Zhang Xinyu, Timothy Putra Prasetio
Abstract: Despite their success in image classification, modern convolutional neural networks (CNNs) exhibit fundamental limitations, including data inefficiency, poor out-of-distribution generalization, and vulnerability to adversarial perturbations. The primate visual system, in contrast, demonstrates superior efficiency and robustness, suggesting that its architectural principles may offer a blueprint for more capable artificial vision systems. This paper introduces Visual Cortex Network (VCNet), a novel neural network architecture whose design is informed by the macro-scale organization of the primate visual cortex. VCNet emulates key biological mechanisms, including hierarchical processing across distinct cortical areas, dual-stream information segregation, and top-down predictive feedback. We evaluate VCNet on two specialized benchmarks: the Spots-10 animal pattern dataset and a light field image classification task. Our results show that VCNet achieves a classification accuracy of 92.1\% on Spots-10 and 74.4\% on the light field dataset, surpassing contemporary models of comparable size. This work demonstrates that integrating neuroscientific principles into network design can lead to more efficient and robust models, providing a promising direction for addressing long-standing challenges in machine learning.
Authors: Meng Zhou, Farzad Khalvati
Abstract: Multimodal medical image fusion integrates complementary information from different imaging modalities to enhance diagnostic accuracy and treatment planning. While deep learning methods have advanced performance, existing approaches face critical limitations: Convolutional Neural Networks (CNNs) excel at local feature extraction but struggle to model global context effectively, while Transformers achieve superior long-range modeling at the cost of quadratic computational complexity, limiting clinical deployment. Recent State Space Models (SSMs) offer a promising alternative, enabling efficient long-range dependency modeling in linear time through selective scan mechanisms. Despite these advances, the extension to 3D volumetric data and the clinical validation of fused images remains underexplored. In this work, we propose ClinicalFMamba, a novel end-to-end CNN-Mamba hybrid architecture that synergistically combines local and global feature modeling for 2D and 3D images. We further design a tri-plane scanning strategy for effectively learning volumetric dependencies in 3D images. Comprehensive evaluations on three datasets demonstrate the superior fusion performance across multiple quantitative metrics while achieving real-time fusion. We further validate the clinical utility of our approach on downstream 2D/3D brain tumor classification tasks, achieving superior performance over baseline methods. Our method establishes a new paradigm for efficient multimodal medical image fusion suitable for real-time clinical deployment.
Authors: Tongxu Zhang, Zhiming Liang, Bei Wang
Abstract: Point clouds have become an increasingly important representation for 3D medical imaging, offering a compact, surface-preserving alternative to traditional voxel or mesh-based approaches. Recent advances in deep learning have enabled rapid progress in extracting, modeling, and analyzing anatomical shapes directly from point cloud data. This paper provides a comprehensive and systematic survey of learning-based shape analysis for medical point clouds, focusing on three fundamental tasks: registration, reconstruction, and variation modeling. We review recent literature from 2021 to 2025, summarize representative methods, datasets, and evaluation metrics, and highlight clinical applications and unique challenges in the medical domain. Key trends include the integration of hybrid representations, large-scale self-supervised models, and generative techniques. We also discuss current limitations, such as data scarcity, inter-patient variability, and the need for interpretable and robust solutions for clinical deployment. Finally, future directions are outlined for advancing point cloud-based shape learning in medical imaging.
Authors: Bo Zhang, JianFei Huo, Zheng Zhang, Wufan Wang, Hui Gao, Xiangyang Gong, Wendong Wang
Abstract: Arbitrary-resolution super-resolution (ARSR) provides crucial flexibility for medical image analysis by adapting to diverse spatial resolutions. However, traditional CNN-based methods are inherently ill-suited for ARSR, as they are typically designed for fixed upsampling factors. While INR-based methods overcome this limitation, they still struggle to effectively process and leverage multi-modal images with varying resolutions and details. In this paper, we propose Nexus-INR, a Diverse Knowledge-guided ARSR framework, which employs varied information and downstream tasks to achieve high-quality, adaptive-resolution medical image super-resolution. Specifically, Nexus-INR contains three key components. A dual-branch encoder with an auxiliary classification task to effectively disentangle shared anatomical structures and modality-specific features; a knowledge distillation module using cross-modal attention that guides low-resolution modality reconstruction with high-resolution reference, enhanced by self-supervised consistency loss; an integrated segmentation module that embeds anatomical semantics to improve both reconstruction quality and downstream segmentation performance. Experiments on the BraTS2020 dataset for both super-resolution and downstream segmentation demonstrate that Nexus-INR outperforms state-of-the-art methods across various metrics.
Authors: Xingjun Ma, Hanxun Huang, Tianwei Song, Ye Sun, Yifeng Gao, Yu-Gang Jiang
Abstract: Large-scale pre-training frameworks like CLIP have revolutionized multimodal learning, but their reliance on web-scraped datasets, frequently containing private user data, raises serious concerns about misuse. Unlearnable Examples (UEs) have emerged as a promising countermeasure against unauthorized model training, employing carefully crafted unlearnable noise to disrupt the learning of meaningful representations from protected data. Current approaches typically generate UEs by jointly optimizing unlearnable noise for both images and their associated text descriptions (or labels). However, this optimization process is often computationally prohibitive for on-device execution, forcing reliance on external third-party services. This creates a fundamental privacy paradox: users must initially expose their data to these very services to achieve protection, thereby compromising privacy in the process. Such a contradiction has severely hindered the development of practical, scalable data protection solutions. To resolve this paradox, we introduce \textbf{Text-to-Unlearnable Example (T2UE)}, a novel framework that enables users to generate UEs using only text descriptions. T2UE circumvents the need for original image data by employing a text-to-image (T2I) model to map text descriptions into the image (noise) space, combined with an error-minimization framework to produce effective unlearnable noise. Extensive experiments show that T2UE-protected data substantially degrades performance in downstream tasks (e.g., cross-modal retrieval) for state-of-the-art models. Notably, the protective effect generalizes across diverse architectures and even to supervised learning settings. Our work demonstrates the feasibility of "zero-contact data protection", where personal data can be safeguarded based solely on their textual descriptions, eliminating the need for direct data exposure.
Authors: Yu Pan, Jiahao Chen, Lin Wang, Bingrong Dai, Yi Du
Abstract: In recent years,Diffusion models have achieved remarkable progress in the field of image generation.However,recent studies have shown that diffusion models are susceptible to backdoor attacks,in which attackers can manipulate the output by injecting covert triggers such as specific visual patterns or textual phrases into the training dataset.Fortunately,with the continuous advancement of defense techniques,defenders have become increasingly capable of identifying and mitigating most backdoor attacks using visual inspection and neural network-based detection methods.However,in this paper,we identify a novel type of backdoor threat that is more lightweight and covert than existing approaches,which we name BadBlocks,requires only about 30\% of the computational resources and 20\% GPU time typically needed by previous backdoor attacks,yet it successfully injects backdoors and evades the most advanced defense frameworks.BadBlocks enables attackers to selectively contaminate specific blocks within the UNet architecture of diffusion models while maintaining normal functionality in the remaining components.Experimental results demonstrate that BadBlocks achieves a high attack success rate (ASR) and low perceptual quality loss (as measured by FID Score),even under extremely constrained computational resources and GPU time.Moreover,BadBlocks is able to bypass existing defense frameworks,especially the attention-based backdoor detection method, highlighting it as a novel and noteworthy threat.Ablation studies further demonstrate that effective backdoor injection does not require fine-tuning the entire network and highlight the pivotal role of certain neural network layers in backdoor mapping.Overall,BadBlocks significantly reduces the barrier to conducting backdoor attacks in all aspects.It enables attackers to inject backdoors into large-scale diffusion models even using consumer-grade GPUs.
Authors: Hengxin Ruan, Qiufan Lin, Shupei Chen, Yang Wang, Wei Zhang
Abstract: Galaxy image translation is an important application in galaxy physics and cosmology. With deep learning-based generative models, image translation has been performed for image generation, data quality enhancement, information extraction, and generalized for other tasks such as deblending and anomaly detection. However, most endeavors on image translation primarily focus on the pixel-level and morphology-level statistics of galaxy images. There is a lack of discussion on the preservation of complex high-order galaxy physical information, which would be more challenging but crucial for studies that rely on high-fidelity image translation. Therefore, we investigated the effectiveness of generative models in preserving high-order physical information (represented by spectroscopic redshift) along with pixel-level and morphology-level information. We tested four representative models, i.e. a Swin Transformer, an SRGAN, a capsule network, and a diffusion model, using the SDSS and CFHTLS galaxy images. We found that these models show different levels of incapabilities in retaining redshift information, even if the global structures of galaxies and morphology-level statistics can be roughly reproduced. In particular, the cross-band peak fluxes of galaxies were found to contain meaningful redshift information, whereas they are subject to noticeable uncertainties in the translation of images, which may substantially be due to the nature of many-to-many mapping. Nonetheless, imperfect translated images may still contain a considerable amount of information and thus hold promise for downstream applications for which high image fidelity is not strongly required. Our work can facilitate further research on how complex physical information is manifested on galaxy images, and it provides implications on the development of image translation models for scientific use.
Authors: Haoran Lin, Wenrui Chen, Xianchi Chen, Fan Yang, Qiang Diao, Wenxin Xie, Sijie Wu, Kailun Yang, Maojun Li, Yaonan Wang
Abstract: Dexterous grasp datasets are vital for embodied intelligence, but mostly emphasize grasp stability, ignoring functional grasps needed for tasks like opening bottle caps or holding cup handles. Most rely on bulky, costly, and hard-to-control high-DOF Shadow Hands. Inspired by the human hand's underactuated mechanism, we establish UniFucGrasp, a universal functional grasp annotation strategy and dataset for multiple dexterous hand types. Based on biomimicry, it maps natural human motions to diverse hand structures and uses geometry-based force closure to ensure functional, stable, human-like grasps. This method supports low-cost, efficient collection of diverse, high-quality functional grasps. Finally, we establish the first multi-hand functional grasp dataset and provide a synthesis model to validate its effectiveness. Experiments on the UFG dataset, IsaacSim, and complex robotic tasks show that our method improves functional manipulation accuracy and grasp stability, enables efficient generalization across diverse robotic hands, and overcomes annotation cost and generalization challenges in dexterous grasping. The project page is at https://haochen611.github.io/UFG.
Authors: Yifei Sun, Zhanghao Chen, Hao Zheng, Yuqing Lu, Lixin Duan, Fenglei Fan, Ahmed Elazab, Xiang Wan, Changmiao Wang, Ruiquan Ge
Abstract: Chest X-Ray (CXR) imaging for pulmonary diagnosis raises significant challenges, primarily because bone structures can obscure critical details necessary for accurate diagnosis. Recent advances in deep learning, particularly with diffusion models, offer significant promise for effectively minimizing the visibility of bone structures in CXR images, thereby improving clarity and diagnostic accuracy. Nevertheless, existing diffusion-based methods for bone suppression in CXR imaging struggle to balance the complete suppression of bones with preserving local texture details. Additionally, their high computational demand and extended processing time hinder their practical use in clinical settings. To address these limitations, we introduce a Global-Local Latent Consistency Model (GL-LCM) architecture. This model combines lung segmentation, dual-path sampling, and global-local fusion, enabling fast high-resolution bone suppression in CXR images. To tackle potential boundary artifacts and detail blurring in local-path sampling, we further propose Local-Enhanced Guidance, which addresses these issues without additional training. Comprehensive experiments on a self-collected dataset SZCH-X-Rays, and the public dataset JSRT, reveal that our GL-LCM delivers superior bone suppression and remarkable computational efficiency, significantly outperforming several competitive methods. Our code is available at https://github.com/diaoquesang/GL-LCM.
Authors: Haotian Wang, Yuzhe Weng, Jun Du, Haoran Xu, Xiaoyan Wu, Shan He, Bing Yin, Cong Liu, Jianqing Gao, Qingfeng Liu
Abstract: The introduction of diffusion models has brought significant advances to the field of audio-driven talking head generation. However, the extremely slow inference speed severely limits the practical implementation of diffusion-based talking head generation models. In this study, we propose READ, the first real-time diffusion-transformer-based talking head generation framework. Our approach first learns a spatiotemporal highly compressed video latent space via a temporal VAE, significantly reducing the token count to accelerate generation. To achieve better audio-visual alignment within this compressed latent space, a pre-trained Speech Autoencoder (SpeechAE) is proposed to generate temporally compressed speech latent codes corresponding to the video latent space. These latent representations are then modeled by a carefully designed Audio-to-Video Diffusion Transformer (A2V-DiT) backbone for efficient talking head synthesis. Furthermore, to ensure temporal consistency and accelerated inference in extended generation, we propose a novel asynchronous noise scheduler (ANS) for both the training and inference process of our framework. The ANS leverages asynchronous add-noise and asynchronous motion-guided generation in the latent space, ensuring consistency in generated video clips. Experimental results demonstrate that READ outperforms state-of-the-art methods by generating competitive talking head videos with significantly reduced runtime, achieving an optimal balance between quality and speed while maintaining robust metric stability in long-time generation.
Authors: Gideon N. L. Rouwendaal, Dani\"el Boeke, Inge L. Cox, Henk G. van der Poel, Margriet C. van Dijk-de Haan, Regina G. H. Beets-Tan, Thierry N. Boellaard, Wilson Silva
Abstract: Accurate preoperative prediction of erectile dysfunction (ED) is important for counseling patients undergoing radical prostatectomy. While clinical features are established predictors, the added value of preoperative MRI remains underexplored. We investigate whether MRI provides additional predictive value for ED at 12 months post-surgery, evaluating four modeling strategies: (1) a clinical-only baseline, representing current state-of-the-art; (2) classical models using handcrafted anatomical features derived from MRI; (3) deep learning models trained directly on MRI slices; and (4) multimodal fusion of imaging and clinical inputs. Imaging-based models (maximum AUC 0.569) slightly outperformed handcrafted anatomical approaches (AUC 0.554) but fell short of the clinical baseline (AUC 0.663). Fusion models offered marginal gains (AUC 0.586) but did not exceed clinical-only performance. SHAP analysis confirmed that clinical features contributed most to predictive performance. Saliency maps from the best-performing imaging model suggested a predominant focus on anatomically plausible regions, such as the prostate and neurovascular bundles. While MRI-based models did not improve predictive performance over clinical features, our findings suggest that they try to capture patterns in relevant anatomical structures and may complement clinical predictors in future multimodal approaches.
Authors: Ana Lawry Aguila, Ayodeji Ijishakin, Juan Eugenio Iglesias, Tomomi Takenaga, Yukihiro Nomura, Takeharu Yoshikawa, Osamu Abe, Shouhei Hanaoka
Abstract: Applying machine learning to real-world medical data, e.g. from hospital archives, has the potential to revolutionize disease detection in brain images. However, detecting pathology in such heterogeneous cohorts is a difficult challenge. Normative modeling, a form of unsupervised anomaly detection, offers a promising approach to studying such cohorts where the ``normal'' behavior is modeled and can be used at subject level to detect deviations relating to disease pathology. Diffusion models have emerged as powerful tools for anomaly detection due to their ability to capture complex data distributions and generate high-quality images. Their performance relies on image restoration; differences between the original and restored images highlight potential abnormalities. However, unlike normative models, these diffusion model approaches do not incorporate clinical information which provides important context to guide the disease detection process. Furthermore, standard approaches often poorly restore healthy regions, resulting in poor reconstructions and suboptimal detection performance. We present CADD, the first conditional diffusion model for normative modeling in 3D images. To guide the healthy restoration process, we propose a novel inference inpainting strategy which balances anomaly removal with retention of subject-specific features. Evaluated on three challenging datasets, including clinical scans, which may have lower contrast, thicker slices, and motion artifacts, CADD achieves state-of-the-art performance in detecting neurological abnormalities in heterogeneous cohorts.
Authors: Wenxuan Shen, Mingjia Wang, Yaochen Wang, Dongping Chen, Junjie Yang, Yao Wan, Weiwei Lin
Abstract: Retrieval-Augmented Generation (RAG) systems using Multimodal Large Language Models (MLLMs) show great promise for complex document understanding, yet their development is critically hampered by inadequate evaluation. Current benchmarks often focus on specific part of document RAG system and use synthetic data with incomplete ground truth and evidence labels, therefore failing to reflect real-world bottlenecks and challenges. To overcome these limitations, we introduce Double-Bench: a new large-scale, multilingual, and multimodal evaluation system that is able to produce fine-grained assessment to each component within document RAG systems. It comprises 3,276 documents (72,880 pages) and 5,168 single- and multi-hop queries across 6 languages and 4 document types with streamlined dynamic update support for potential data contamination issues. Queries are grounded in exhaustively scanned evidence pages and verified by human experts to ensure maximum quality and completeness. Our comprehensive experiments across 9 state-of-the-art embedding models, 4 MLLMs and 4 end-to-end document RAG frameworks demonstrate the gap between text and visual embedding models is narrowing, highlighting the need in building stronger document retrieval models. Our findings also reveal the over-confidence dilemma within current document RAG frameworks that tend to provide answer even without evidence support. We hope our fully open-source Double-Bench provide a rigorous foundation for future research in advanced document RAG systems. We plan to retrieve timely corpus and release new benchmarks on an annual basis.
Authors: Akshay L Chandra, Iman Nematollahi, Chenguang Huang, Tim Welschehold, Wolfram Burgard, Abhinav Valada
Abstract: Fine-tuning diffusion policies with reinforcement learning (RL) presents significant challenges. The long denoising sequence for each action prediction impedes effective reward propagation. Moreover, standard RL methods require millions of real-world interactions, posing a major bottleneck for practical fine-tuning. Although prior work frames the denoising process in diffusion policies as a Markov Decision Process to enable RL-based updates, its strong dependence on environment interaction remains highly inefficient. To bridge this gap, we introduce DiWA, a novel framework that leverages a world model for fine-tuning diffusion-based robotic skills entirely offline with reinforcement learning. Unlike model-free approaches that require millions of environment interactions to fine-tune a repertoire of robot skills, DiWA achieves effective adaptation using a world model trained once on a few hundred thousand offline play interactions. This results in dramatically improved sample efficiency, making the approach significantly more practical and safer for real-world robot learning. On the challenging CALVIN benchmark, DiWA improves performance across eight tasks using only offline adaptation, while requiring orders of magnitude fewer physical interactions than model-free baselines. To our knowledge, this is the first demonstration of fine-tuning diffusion policies for real-world robotic skills using an offline world model. We make the code publicly available at https://diwa.cs.uni-freiburg.de.
Authors: Xinyu Wang, Yue Zhang, Liqiang Jing
Abstract: Sarcasm is a complex linguistic phenomenon that involves a disparity between literal and intended meanings, making it challenging for sentiment analysis and other emotion-sensitive tasks. While traditional sarcasm detection methods primarily focus on text, recent approaches have incorporated multimodal information. However, the application of Large Visual Language Models (LVLMs) in Multimodal Sarcasm Analysis (MSA) remains underexplored. In this paper, we evaluate LVLMs in MSA tasks, specifically focusing on Multimodal Sarcasm Detection and Multimodal Sarcasm Explanation. Through comprehensive experiments, we identify key limitations, such as insufficient visual understanding and a lack of conceptual knowledge. To address these issues, we propose a training-free framework that integrates in-depth object extraction and external conceptual knowledge to improve the model's ability to interpret and explain sarcasm in multimodal contexts. The experimental results on multiple models show the effectiveness of our proposed framework. The code is available at https://github.com/cp-cp/LVLM-MSA.
Authors: Prateek Mittal, Puneet Goyal, Joohi Chauhan
Abstract: This study introduces a novel multimodal food recognition framework that effectively combines visual and textual modalities to enhance classification accuracy and robustness. The proposed approach employs a dynamic multimodal fusion strategy that adaptively integrates features from unimodal visual inputs and complementary textual metadata. This fusion mechanism is designed to maximize the use of informative content, while mitigating the adverse impact of missing or inconsistent modality data. The framework was rigorously evaluated on the UPMC Food-101 dataset and achieved unimodal classification accuracies of 73.60% for images and 88.84% for text. When both modalities were fused, the model achieved an accuracy of 97.84%, outperforming several state-of-the-art methods. Extensive experimental analysis demonstrated the robustness, adaptability, and computational efficiency of the proposed settings, highlighting its practical applicability to real-world multimodal food-recognition scenarios.
Authors: Md. Ismail Hossain, M M Lutfe Elahi, Sameera Ramasinghe, Ali Cheraghian, Fuad Rahman, Nabeel Mohammed, Shafin Rahman
Abstract: In the knowledge distillation literature, feature-based methods have dominated due to their ability to effectively tap into extensive teacher models. In contrast, logit-based approaches, which aim to distill "dark knowledge" from teachers, typically exhibit inferior performance compared to feature-based methods. To bridge this gap, we present LumiNet, a novel knowledge distillation algorithm designed to enhance logit-based distillation. We introduce the concept of "perception", aiming to calibrate logits based on the model's representation capability. This concept addresses overconfidence issues in the logit-based distillation method while also introducing a novel method to distill knowledge from the teacher. It reconstructs the logits of a sample/instances by considering relationships with other samples in the batch. LumiNet excels on benchmarks like CIFAR-100, ImageNet, and MSCOCO, outperforming the leading feature-based methods, e.g., compared to KD with ResNet18 and MobileNetV2 on ImageNet, it shows improvements of 1.5% and 2.05%, respectively. Codes are available at https://github.com/ismail31416/LumiNet.
Authors: Stefan Kolek, Aditya Chattopadhyay, Kwan Ho Ryan Chan, Hector Andrade-Loarca, Gitta Kutyniok, R\'ene Vidal
Abstract: Information Pursuit (IP) is an explainable prediction algorithm that greedily selects a sequence of interpretable queries about the data in order of information gain, updating its posterior at each step based on observed query-answer pairs. The standard paradigm uses hand-crafted dictionaries of potential data queries curated by a domain expert or a large language model after a human prompt. However, in practice, hand-crafted dictionaries are limited by the expertise of the curator and the heuristics of prompt engineering. This paper introduces a novel approach: learning a dictionary of interpretable queries directly from the dataset. Our query dictionary learning problem is formulated as an optimization problem by augmenting IP's variational formulation with learnable dictionary parameters. To formulate learnable and interpretable queries, we leverage the latent space of large vision and language models like CLIP. To solve the optimization problem, we propose a new query dictionary learning algorithm inspired by classical sparse dictionary learning. Our experiments demonstrate that learned dictionaries significantly outperform hand-crafted dictionaries generated with large language models.
Authors: Ke Zou, Yang Bai, Bo Liu, Yidi Chen, Zhihao Chen, Yang Zhou, Xuedong Yuan, Meng Wang, Xiaojing Shen, Xiaochun Cao, Yih Chung Tham, Huazhu Fu
Abstract: Medical phrase grounding is crucial for identifying relevant regions in medical images based on phrase queries, facilitating accurate image analysis and diagnosis. However, current methods rely on manual extraction of key phrases from medical reports, reducing efficiency and increasing the workload for clinicians. Additionally, the lack of model confidence estimation limits clinical trust and usability. In this paper, we introduce a novel task called Medical Report Grounding (MRG), which aims to directly identify diagnostic phrases and their corresponding grounding boxes from medical reports in an end-to-end manner. To address this challenge, we propose uMedGround, a robust and reliable framework that leverages a multimodal large language model to predict diagnostic phrases by embedding a unique token,
Authors: Ling-Qi Zhang, Zahra Kadkhodaie, Eero P. Simoncelli, David H. Brainard
Abstract: We examine the problem of selecting a small set of linear measurements for reconstructing high-dimensional signals. Well-established methods for optimizing such measurements include principal component analysis (PCA), independent component analysis (ICA) and compressed sensing (CS) based on random projections, all of which rely on axis- or subspace-aligned statistical characterization of the signal source. However, many naturally occurring signals, including photographic images, contain richer statistical structure. To exploit such structure, we introduce a general method for obtaining an optimized set of linear measurements for efficient image reconstruction, where the signal statistics are expressed by the prior implicit in a neural network trained to perform denoising (known as a "diffusion model"). We demonstrate that the optimal measurements derived for two natural image datasets differ from those of PCA, ICA, or CS, and result in substantially lower mean squared reconstruction error. Interestingly, the marginal distributions of the measurement values are asymmetrical (skewed), substantially more so than those of previous methods. We also find that optimizing with respect to perceptual loss, as quantified by structural similarity (SSIM), leads to measurements different from those obtained when optimizing for MSE. Our results highlight the importance of incorporating the specific statistical regularities of natural signals when designing effective linear measurements.
Authors: Xiaozhe Zhang, Fengying Xie, Haidong Ding, Linpeng Pan, Zhenwei Shi
Abstract: Haze removal aims to restore a clear image from a hazy input. Existing methods achieve notable success by specializing in either short-range dependencies to preserve local details or long-range dependencies to capture global context. Given the complementary strengths of both, a natural progression is to explicitly integrate them within a unified framework and enable their reasonable aggregation. However, this integration remains underexplored. In this paper, we propose DehazeMatic, which simultaneously and explicitly captures both short- and long-range dependencies through a dual-stream design. To optimize the contribution of dependencies at varying ranges, we conduct extensive experiments to identify key influencing factors and find that an effective aggregation mechanism should be guided by the joint consideration of haze density and semantic information. Building on these insights, we introduce the CLIP-enhanced Dual-path Aggregator, which not only enables the generation of fine-grained haze density maps for the first time, but also produces semantic maps within a shared backbone, ultimately leveraging both to instruct the aggregation process. Extensive experiments demonstrate that DehazeMatic outperforms state-of-the-art methods across multiple benchmarks.
Authors: Jiawei Lian, Shaohui Mei, Xiaofei Wang, Yi Wang, Lefan Wang, Yingjie Lu, Mingyang Ma, Lap-Pui Chau
Abstract: It has been widely substantiated that deep neural networks (DNNs) are susceptible and vulnerable to adversarial perturbations. Existing studies mainly focus on performing attacks by corrupting targeted objects (physical attack) or images (digital attack), which is intuitively acceptable and understandable in terms of the attack's effectiveness. In contrast, our focus lies in conducting background adversarial attacks in both digital and physical domains, without causing any disruptions to the targeted objects themselves. Specifically, an effective background adversarial attack framework is proposed to attack anything, by which the attack efficacy generalizes well between diverse objects, models, and tasks. Technically, we approach the background adversarial attack as an iterative optimization problem, analogous to the process of DNN learning. Besides, we offer a theoretical demonstration of its convergence under a set of mild but sufficient conditions. To strengthen the attack efficacy and transferability, we propose a new ensemble strategy tailored for adversarial perturbations and introduce an improved smooth constraint for the seamless connection of integrated perturbations. We conduct comprehensive and rigorous experiments in both digital and physical domains across various objects, models, and tasks, demonstrating the effectiveness of attacking anything of the proposed method. The findings of this research substantiate the significant discrepancy between human and machine vision on the value of background variations, which play a far more critical role than previously recognized, necessitating a reevaluation of the robustness and reliability of DNNs. The code will be publicly available at https://github.com/JiaweiLian/Attack_Anything
Authors: Shuai Zhang, Guanjun Wu, Zhoufeng Xie, Xinggang Wang, Bin Feng, Wenyu Liu
Abstract: Reconstructing objects and extracting high-quality surfaces play a vital role in the real world. Current 4D representations show the ability to render high-quality novel views for dynamic objects, but cannot reconstruct high-quality meshes due to their implicit or geometrically inaccurate representations. In this paper, we propose a novel representation that can reconstruct accurate meshes from sparse image input, named Dynamic 2D Gaussians (D-2DGS). We adopt 2D Gaussians for basic geometry representation and use sparse-controlled points to capture the 2D Gaussian's deformation. By extracting the object mask from the rendered high-quality image and masking the rendered depth map, we remove floaters that are prone to occur during reconstruction and can extract high-quality dynamic mesh sequences of dynamic objects. Experiments demonstrate that our D-2DGS is outstanding in reconstructing detailed and smooth high-quality meshes from sparse inputs. The code is available at https://github.com/hustvl/Dynamic-2DGS.
Authors: Zhenyu Yan, Jian Wang, Aoqiang Wang, Yuhan Li, Wenxiang Shang, Ran Lin
Abstract: In image editing tasks, high-quality text editing capabilities can significantly reduce both human and material resource costs. Existing methods, however, face significant limitations in terms of stroke accuracy for complex text and controllability of generated text styles. To address these challenges, we propose TextMaster, a solution capable of accurately editing text across various scenarios and image regions, while ensuring proper layout and controllable text style. Our method enhances the accuracy and fidelity of text rendering by incorporating high-resolution standard glyph information and applying perceptual loss within the text editing region. Additionally, we leverage an attention mechanism to compute intermediate layer bounding box regression loss for each character, enabling the model to learn text layout across varying contexts. Furthermore, we propose a novel style injection technique that enables controllable style transfer for the injected text. Through comprehensive experiments, we demonstrate the state-of-the-art performance of our method.
Authors: Ant Duru, Alptekin Temizel
Abstract: Data augmentation is a critical component of deep learning pipelines, enhancing model generalization by increasing dataset diversity. Traditional augmentation strategies rely on manually designed transformations, stochastic sampling, or automated search-based approaches. Although automated methods improve performance, they often require extensive computational resources and are specifically designed for certain datasets. In this work, we propose a Large Language Model (LLM)-guided augmentation optimization strategy that refines augmentation policies based on model performance feedback. We propose two approaches: (1) LLM-Guided Augmentation Policy Optimization, where augmentation policies selected by LLM are refined iteratively across training cycles, and (2) Adaptive LLM-Guided Augmentation Policy Optimization, which adjusts policies at each iteration based on performance metrics. This in-training approach eliminates the need for full model retraining before getting LLM feedback, reducing computational costs while increasing performance. Our methodology employs an LLM to dynamically select augmentation transformations based on dataset characteristics, model architecture, and prior training performance. Leveraging LLMs' contextual knowledge, especially in domain-specific tasks like medical imaging, our method selects augmentations tailored to dataset characteristics and model performance. Experiments across domain-specific image classification datasets show consistent accuracy improvements over traditional methods.
Authors: Ruofan Wang, Juncheng Li, Yixu Wang, Bo Wang, Xiaosen Wang, Yan Teng, Yingchun Wang, Xingjun Ma, Yu-Gang Jiang
Abstract: As large Vision-Language Models (VLMs) gain prominence, ensuring their safe deployment has become critical. Recent studies have explored VLM robustness against jailbreak attacks-techniques that exploit model vulnerabilities to elicit harmful outputs. However, the limited availability of diverse multimodal data has constrained current approaches to rely heavily on adversarial or manually crafted images derived from harmful text datasets, which often lack effectiveness and diversity across different contexts. In this paper, we propose IDEATOR, a novel jailbreak method that autonomously generates malicious image-text pairs for black-box jailbreak attacks. IDEATOR is grounded in the insight that VLMs themselves could serve as powerful red team models for generating multimodal jailbreak prompts. Specifically, IDEATOR leverages a VLM to create targeted jailbreak texts and pairs them with jailbreak images generated by a state-of-the-art diffusion model. Extensive experiments demonstrate IDEATOR's high effectiveness and transferability, achieving a 94% attack success rate (ASR) in jailbreaking MiniGPT-4 with an average of only 5.34 queries, and high ASRs of 82%, 88%, and 75% when transferred to LLaVA, InstructBLIP, and Chameleon, respectively. Building on IDEATOR's strong transferability and automated process, we introduce the VLJailbreakBench, a safety benchmark comprising 3,654 multimodal jailbreak samples. Our benchmark results on 11 recently released VLMs reveal significant gaps in safety alignment. For instance, our challenge set achieves ASRs of 46.31% on GPT-4o and 19.65% on Claude-3.5-Sonnet, underscoring the urgent need for stronger defenses.
Authors: Jingwei Xu, Chenyu Wang, Zibo Zhao, Wen Liu, Yi Ma, Shenghua Gao
Abstract: This paper aims to design a unified Computer-Aided Design (CAD) generation system that can easily generate CAD models based on the user's inputs in the form of textual description, images, point clouds, or even a combination of them. Towards this goal, we introduce the CAD-MLLM, the first system capable of generating parametric CAD models conditioned on the multimodal input. Specifically, within the CAD-MLLM framework, we leverage the command sequences of CAD models and then employ advanced large language models (LLMs) to align the feature space across these diverse multi-modalities data and CAD models' vectorized representations. To facilitate the model training, we design a comprehensive data construction and annotation pipeline that equips each CAD model with corresponding multimodal data. Our resulting dataset, named Omni-CAD, is the first multimodal CAD dataset that contains textual description, multi-view images, points, and command sequence for each CAD model. It contains approximately 450K instances and their CAD construction sequences. To thoroughly evaluate the quality of our generated CAD models, we go beyond current evaluation metrics that focus on reconstruction quality by introducing additional metrics that assess topology quality and surface enclosure extent. Extensive experimental results demonstrate that CAD-MLLM significantly outperforms existing conditional generative methods and remains highly robust to noises and missing points. The project page and more visualizations can be found at: https://cad-mllm.github.io/
Authors: Zhanqiang Guo, Jiamin Wu, Yonghao Song, Jiahui Bu, Weijian Mai, Qihao Zheng, Wanli Ouyang, Chunfeng Song
Abstract: Human's perception of the visual world is shaped by the stereo processing of 3D information. Understanding how the brain perceives and processes 3D visual stimuli in the real world has been a longstanding endeavor in neuroscience. Towards this goal, we introduce a new neuroscience task: decoding 3D visual perception from EEG signals, a neuroimaging technique that enables real-time monitoring of neural dynamics enriched with complex visual cues. To provide the essential benchmark, we first present EEG-3D, a pioneering dataset featuring multimodal analysis data and extensive EEG recordings from 12 subjects viewing 72 categories of 3D objects rendered in both videos and images. Furthermore, we propose Neuro-3D, a 3D visual decoding framework based on EEG signals. This framework adaptively integrates EEG features derived from static and dynamic stimuli to learn complementary and robust neural representations, which are subsequently utilized to recover both the shape and color of 3D objects through the proposed diffusion-based colored point cloud decoder. To the best of our knowledge, we are the first to explore EEG-based 3D visual decoding. Experiments indicate that Neuro-3D not only reconstructs colored 3D objects with high fidelity, but also learns effective neural representations that enable insightful brain region analysis. The dataset and associated code will be made publicly available.
Authors: Pura Peetathawatchai, Wei-Ning Chen, Berivan Isik, Sanmi Koyejo, Albert No
Abstract: Personalizing large-scale diffusion models poses serious privacy risks, especially when adapting to small, sensitive datasets. A common approach is to fine-tune the model using differentially private stochastic gradient descent (DP-SGD), but this suffers from severe utility degradation due to the high noise needed for privacy, particularly in the small data regime. We propose an alternative that leverages Textual Inversion (TI), which learns an embedding vector for an image or set of images, to enable adaptation under differential privacy (DP) constraints. Our approach, Differentially Private Aggregation via Textual Inversion (DPAgg-TI), adds calibrated noise to the aggregation of per-image embeddings to ensure formal DP guarantees while preserving high output fidelity. We show that DPAgg-TI outperforms DP-SGD finetuning in both utility and robustness under the same privacy budget, achieving results closely matching the non-private baseline on style adaptation tasks using private artwork from a single artist and Paris 2024 Olympic pictograms. In contrast, DP-SGD fails to generate meaningful outputs in this setting.
Authors: Woong Oh Cho, In Cho, Seoha Kim, Jeongmin Bae, Youngjung Uh, Seon Joo Kim
Abstract: Modeling dynamic scenes through 4D Gaussians offers high visual fidelity and fast rendering speeds, but comes with significant storage overhead. Recent approaches mitigate this cost by aggressively reducing the number of Gaussians. However, this inevitably removes Gaussians essential for high-quality rendering, leading to severe degradation in dynamic regions. In this paper, we introduce a novel 4D anchor-based framework that tackles the storage cost in different perspective. Rather than reducing the number of Gaussians, our method retains a sufficient quantity to accurately model dynamic contents, while compressing them into compact, grid-aligned 4D anchor features. Each anchor is processed by an MLP to spawn a set of neural 4D Gaussians, which represent a local spatiotemporal region. We design these neural 4D Gaussians to capture temporal changes with minimal parameters, making them well-suited for the MLP-based spawning. Moreover, we introduce a dynamic-aware anchor growing strategy to effectively assign additional anchors to under-reconstructed dynamic regions. Our method adjusts the accumulated gradients with Gaussians' temporal coverage, significantly improving reconstruction quality in dynamic regions. Experimental results highlight that our method achieves state-of-the-art visual quality in dynamic regions, outperforming all baselines by a large margin with practical storage costs.
Authors: Yiming Wu, Zhenghao Chen, Huan Wang, Dong Xu
Abstract: The high computational cost and slow inference time are major obstacles to deploying Video Diffusion Models (VDMs). To overcome this, we introduce a new Video Diffusion Model Compression approach using individual content and motion dynamics preserved pruning and consistency loss. First, we empirically observe that deeper VDM layers are crucial for maintaining the quality of \textbf{motion dynamics} (\textit{e.g.,} coherence of the entire video), while shallower layers are more focused on \textbf{individual content} (\textit{e.g.,} individual frames). Therefore, we prune redundant blocks from the shallower layers while preserving more of the deeper layers, resulting in a lightweight VDM variant called VDMini. Moreover, we propose an \textbf{Individual Content and Motion Dynamics (ICMD)} Consistency Loss to gain comparable generation performance as larger VDM to VDMini. In particular, we first use the Individual Content Distillation (ICD) Loss to preserve the consistency in the features of each generated frame between the teacher and student models. Next, we introduce a Multi-frame Content Adversarial (MCA) Loss to enhance the motion dynamics across the generated video as a whole. This method significantly accelerates inference time while maintaining high-quality video generation. Extensive experiments demonstrate the effectiveness of our VDMini on two important video generation tasks, Text-to-Video (T2V) and Image-to-Video (I2V), where we respectively achieve an average 2.5 $\times$, 1.4 $\times$, and 1.25 $\times$ speed up for the I2V method SF-V, the T2V method T2V-Turbo-v2, and the T2V method HunyuanVideo, while maintaining the quality of the generated videos on several benchmarks including UCF101, VBench-T2V, and VBench-I2V.
Authors: Luca Barsellotti, Lorenzo Bianchi, Nicola Messina, Fabio Carrara, Marcella Cornia, Lorenzo Baraldi, Fabrizio Falchi, Rita Cucchiara
Abstract: Open-Vocabulary Segmentation (OVS) aims at segmenting images from free-form textual concepts without predefined training classes. While existing vision-language models such as CLIP can generate segmentation masks by leveraging coarse spatial information from Vision Transformers, they face challenges in spatial localization due to their global alignment of image and text features. Conversely, self-supervised visual models like DINO excel in fine-grained visual encoding but lack integration with language. To bridge this gap, we present Talk2DINO, a novel hybrid approach that combines the spatial accuracy of DINOv2 with the language understanding of CLIP. Our approach aligns the textual embeddings of CLIP to the patch-level features of DINOv2 through a learned mapping function without the need to fine-tune the underlying backbones. At training time, we exploit the attention maps of DINOv2 to selectively align local visual patches with textual embeddings. We show that the powerful semantic and localization abilities of Talk2DINO can enhance the segmentation process, resulting in more natural and less noisy segmentations, and that our approach can also effectively distinguish foreground objects from the background. Experimental results demonstrate that Talk2DINO achieves state-of-the-art performance across several unsupervised OVS benchmarks. Source code and models are publicly available at: https://lorebianchi98.github.io/Talk2DINO/.
Authors: Yuci Liang, Xinheng Lyu, Meidan Ding, Wenting Chen, Jipeng Zhang, Yuexiang Ren, Xiangjian He, Song Wu, Sen Yang, Xiyue Wang, Xiaohan Xing, Linlin Shen
Abstract: Recent advancements in computational pathology have produced patch-level Multi-modal Large Language Models (MLLMs), but these models are limited by their inability to analyze whole slide images (WSIs) comprehensively and their tendency to bypass crucial morphological features that pathologists rely on for diagnosis. To address these challenges, we first introduce WSI-Bench, a large-scale morphology-aware benchmark containing 180k VQA pairs from 9,850 WSIs across 30 cancer types, designed to evaluate MLLMs' understanding of morphological characteristics crucial for accurate diagnosis. Building upon this benchmark, we present WSI-LLaVA, a novel framework for gigapixel WSI understanding that employs a three-stage training approach: WSI-text alignment, feature space alignment, and task-specific instruction tuning. To better assess model performance in pathological contexts, we develop two specialized WSI metrics: WSI-Precision and WSI-Relevance. Experimental results demonstrate that WSI-LLaVA outperforms existing models across all capability dimensions, with a significant improvement in morphological analysis, establishing a clear correlation between morphological understanding and diagnostic accuracy.
Authors: Dmitrii Torbunov, Yihui Ren, Animesh Ghose, Odera Dim, Yonggang Cui
Abstract: Event-based cameras (EBCs) have emerged as a bio-inspired alternative to traditional cameras, offering advantages in power efficiency, temporal resolution, and high dynamic range. However, the development of image analysis methods for EBCs is challenging due to the sparse and asynchronous nature of the data. This work addresses the problem of object detection for EBC cameras. The current approaches to EBC object detection focus on constructing complex data representations and rely on specialized architectures. We introduce I2EvDet (Image-to-Event Detection), a novel adaptation framework that bridges mainstream object detection with temporal event data processing. First, we demonstrate that a Real-Time DEtection TRansformer, or RT-DETR, a state-of-the-art natural image detector, trained on a simple image-like representation of the EBC data achieves performance comparable to specialized EBC methods. Next, as part of our framework, we develop an efficient adaptation technique that transforms image-based detectors into event-based detection models by modifying their frozen latent representation space through minimal architectural additions. The resulting EvRT-DETR model reaches state-of-the-art performance on the standard benchmark datasets Gen1 (mAP $+2.3$) and 1Mpx/Gen4 (mAP $+1.4$). These results demonstrate a fundamentally new approach to EBC object detection through principled adaptation of mainstream architectures, offering an efficient alternative with potential applications to other temporal visual domains. The code is available at: https://github.com/realtime-intelligence/evrt-detr
Authors: Anand Kumar, Jiteng Mu, Nuno Vasconcelos
Abstract: Text-to-image (T2I) models have recently gained widespread adoption. This has spurred concerns about safeguarding intellectual property rights and an increasing demand for mechanisms that prevent the generation of specific artistic styles. Existing methods for style extraction typically necessitate the collection of custom datasets and the training of specialized models. This, however, is resource-intensive, time-consuming, and often impractical for real-time applications. We present a novel, training-free framework to solve the style attribution problem, using the features produced by a diffusion model alone, without any external modules or retraining. This is denoted as Introspective Style attribution (IntroStyle) and is shown to have superior performance to state-of-the-art models for style attribution. We also introduce a synthetic dataset of Artistic Style Split (ArtSplit) to isolate artistic style and evaluate fine-grained style attribution performance. Our experimental results on WikiArt and DomainNet datasets show that \ours is robust to the dynamic nature of artistic styles, outperforming existing methods by a wide margin.
Authors: Faraz Waseem, Muhammad Shahzad
Abstract: An image may convey a thousand words, but a video composed of hundreds or thousands of image frames tells a more intricate story. Despite significant progress in multimodal large language models (MLLMs), generating extended videos remains a formidable challenge. As of this writing, OpenAI's Sora, the current state-of-the-art system, is still limited to producing videos that are up to one minute in length. This limitation stems from the complexity of long video generation, which requires more than generative AI techniques for approximating density functions essential aspects such as planning, story development, and maintaining spatial and temporal consistency present additional hurdles. Integrating generative AI with a divide-and-conquer approach could improve scalability for longer videos while offering greater control. In this survey, we examine the current landscape of long video generation, covering foundational techniques like GANs and diffusion models, video generation strategies, large-scale training datasets, quality metrics for evaluating long videos, and future research areas to address the limitations of the existing video generation capabilities. We believe it would serve as a comprehensive foundation, offering extensive information to guide future advancements and research in the field of long video generation.
Authors: Zhipeng Deng, Zhe Xu, Tsuyoshi Isshiki, Yefeng Zheng
Abstract: Medical image segmentation is challenging due to the diversity of medical images and the lack of labeled data, which motivates recent developments in federated semi-supervised learning (FSSL) to leverage a large amount of unlabeled data from multiple centers for model training without sharing raw data. However, what remains under-explored in FSSL is the domain shift problem which may cause suboptimal model aggregation and low effectivity of the utilization of unlabeled data, eventually leading to unsatisfactory performance in unseen domains. In this paper, we explore this previously ignored scenario, namely domain generalized federated semi-supervised learning (FedSemiDG), which aims to learn a model in a distributed manner from multiple domains with limited labeled data and abundant unlabeled data such that the model can generalize well to unseen domains. We present a novel framework, Federated Generalization-Aware SemiSupervised Learning (FGASL), to address the challenges in FedSemiDG by effectively tackling critical issues at both global and local levels. Globally, we introduce Generalization-Aware Aggregation (GAA), assigning adaptive weights to local models based on their generalization performance. Locally, we use a Dual-Teacher Adaptive Pseudo Label Refinement (DR) strategy to combine global and domain-specific knowledge, generating more reliable pseudo labels. Additionally, Perturbation-Invariant Alignment (PIA) enforces feature consistency under perturbations, promoting domain-invariant learning. Extensive experiments on four medical segmentation tasks (cardiac MRI, spine MRI, bladder cancer MRI and colorectal polyp) demonstrate that our method significantly outperforms state-of-the-art FSSL and domain generalization approaches, achieving robust generalization on unseen domains.
Authors: Alexis Roger, Prateek Humane, Daniel Z. Kaplan, Kshitij Gupta, Qi Sun, George Adamopoulos, Jonathan Siu Chi Lim, Quentin Anthony, Edwin Fennell, Irina Rish
Abstract: The proliferation of Vision-Language Models (VLMs) in the past several years calls for rigorous and comprehensive evaluation methods and benchmarks. This work analyzes existing VLM evaluation techniques, including automated metrics, AI-based assessments, and human evaluations across diverse tasks. We first introduce Robin - a novel suite of VLMs that we built by combining Large Language Models (LLMs) and Vision Encoders (VEs) at multiple scales, and use Robin to identify shortcomings of current evaluation approaches across scales. Next, to overcome the identified limitations, we introduce CHIRP - a new long form response benchmark we developed for more robust and complete VLM evaluation. We provide open access to the Robin training code, model suite, and CHIRP benchmark to promote reproducibility and advance VLM research.
Authors: Tiancheng Gu, Kaicheng Yang, Chaoyi Zhang, Yin Xie, Xiang An, Ziyong Feng, Dongnan Liu, Weidong Cai, Jiankang Deng
Abstract: After pre-training on extensive image-text pairs, Contrastive Language-Image Pre-training (CLIP) demonstrates promising performance on a wide variety of benchmarks. However, a substantial volume of multimodal interleaved documents remains underutilized for contrastive vision-language representation learning. To fully leverage these unpaired documents, we initially establish a Real-World Data Extraction pipeline to extract high-quality images and texts. Then we design a hierarchical retrieval method to efficiently associate each image with multiple semantically relevant realistic texts. To further enhance fine-grained visual information, we propose an image semantic augmented generation module for synthetic text production. Furthermore, we employ a semantic balance sampling strategy to improve dataset diversity, enabling better learning of long-tail concepts. Based on these innovations, we construct RealSyn, a dataset combining realistic and synthetic texts, available in three scales: 15M, 30M, and 100M. We compare our dataset with other widely used datasets of equivalent scale for CLIP training. Models pre-trained on RealSyn consistently achieve state-of-the-art performance across various downstream tasks, including linear probe, zero-shot transfer, zero-shot robustness, and zero-shot retrieval. Furthermore, extensive experiments confirm that RealSyn significantly enhances contrastive vision-language representation learning and demonstrates robust scalability. To facilitate future research, the RealSyn dataset and pretrained model weights are released at https://github.com/deepglint/RealSyn.
Authors: Cong-Duy Nguyen, Xiaobao Wu, Duc Anh Vu, Shuai Zhao, Thong Nguyen, Anh Tuan Luu
Abstract: Large Vision-Language Models (LVLMs) have demonstrated impressive multimodal reasoning capabilities, but they remain susceptible to hallucination, particularly object hallucination where non-existent objects or incorrect attributes are fabricated in generated descriptions. Existing detection methods achieve strong performance but rely heavily on expensive API calls and iterative LVLM-based validation, making them impractical for large-scale or offline use. To address these limitations, we propose CutPaste\&Find, a lightweight and training-free framework for detecting hallucinations in LVLM-generated outputs. Our approach leverages off-the-shelf visual and linguistic modules to perform multi-step verification efficiently without requiring LVLM inference. At the core of our framework is a Visual-aid Knowledge Base that encodes rich entity-attribute relationships and associated image representations. We introduce a scaling factor to refine similarity scores, mitigating the issue of suboptimal alignment values even for ground-truth image-text pairs. Comprehensive evaluations on benchmark datasets, including POPE and R-Bench, demonstrate that CutPaste\&Find achieves competitive hallucination detection performance while being significantly more efficient and cost-effective than previous methods.
Authors: Yueyang Li, Lei Chen, Wenhao Dong, Shengyu Gong, Zijian Kang, Boyang Wei, Weiming Zeng, Hongjie Yan, Lingbin Bian, Zhiguo Zhang, Wai Ting Siok, Nizhuan Wang
Abstract: Developing interpretable models for neurodevelopmental disorders (NDDs) diagnosis presents significant challenges in effectively encoding, decoding, and integrating multimodal neuroimaging data. While many existing machine learning approaches have shown promise in brain network analysis, they typically suffer from limited interpretability, particularly in extracting meaningful biomarkers from functional magnetic resonance imaging (fMRI) data and establishing clear relationships between imaging features and demographic characteristics. Besides, current graph neural network methodologies face limitations in capturing both local and global functional connectivity patterns while simultaneously achieving theoretically principled multimodal data fusion. To address these challenges, we propose the Interpretable Information Bottleneck Heterogeneous Graph Neural Network (I2B-HGNN), a unified framework that applies information bottleneck principles to guide both brain connectivity modeling and cross-modal feature integration. This framework comprises two complementary components. The first is the Information Bottleneck Graph Transformer (IBGraphFormer), which combines transformer-based global attention mechanisms with graph neural networks through information bottleneck-guided pooling to identify sufficient biomarkers. The second is the Information Bottleneck Heterogeneous Graph Attention Network (IB-HGAN), which employs meta-path-based heterogeneous graph learning with structural consistency constraints to achieve interpretable fusion of neuroimaging and demographic data. The experimental results demonstrate that I2B-HGNN achieves superior performance in diagnosing NDDs, exhibiting both high classification accuracy and the ability to provide interpretable biomarker identification while effectively analyzing non-imaging data.
Authors: Nathan Drenkow, Mathias Unberath
Abstract: Image quality plays an important role in the performance of deep neural networks (DNNs) that have been widely shown to exhibit sensitivity to changes in imaging conditions. Conventional image quality assessment (IQA) seeks to measure and align quality relative to human perceptual judgments, but we often need a metric that is not only sensitive to imaging conditions but also well-aligned with DNN sensitivities. We first ask whether conventional IQA metrics are also informative of DNN performance. We show theoretically and empirically that conventional IQA metrics are weak predictors of DNN performance for image classification. Using our causal framework, we then develop metrics that exhibit strong correlation with DNN performance, thus enabling us to effectively estimate the quality distribution of large image datasets relative to targeted vision tasks.
Authors: Seungju Cho, Hongsin Lee, Changick Kim
Abstract: Adversarial training significantly enhances adversarial robustness, yet superior performance is predominantly achieved on balanced datasets. Addressing adversarial robustness in the context of unbalanced or long-tailed distributions is considerably more challenging, mainly due to the scarcity of tail data instances. Previous research on adversarial robustness within long-tailed distributions has primarily focused on combining traditional long-tailed natural training with existing adversarial robustness methods. In this study, we provide an in-depth analysis for the challenge that adversarial training struggles to achieve high performance on tail classes in long-tailed distributions. Furthermore, we propose a simple yet effective solution to advance adversarial robustness on long-tailed distributions through a novel self-distillation technique. Specifically, this approach leverages a balanced self-teacher model, which is trained using a balanced dataset sampled from the original long-tailed dataset. Our extensive experiments demonstrate state-of-the-art performance in both clean and robust accuracy for long-tailed adversarial robustness, with significant improvements in tail class performance on various datasets. We improve the accuracy against PGD attacks for tail classes by 20.3, 7.1, and 3.8 percentage points on CIFAR-10, CIFAR-100, and Tiny-ImageNet, respectively, while achieving the highest robust accuracy.
Authors: Zhen Zou, Feng Zhao
Abstract: Diffusion Transformer (DiT) has exhibited impressive generation capabilities but faces great challenges due to its high computational complexity. To address this issue, various methods, notably feature caching, have been introduced. However, these approaches focus on aligning non-cache diffusion without analyzing why caching damage the generation processes. In this paper, we first confirm that the cache greatly amplifies the exposure bias, resulting in a decline in the generation quality. However, directly applying noise scaling is challenging for this issue due to the non-smoothness of exposure bias. We found that this phenomenon stems from the mismatch between its frequency response characteristics and the simple cache of Attention and MLP. Since these two components exhibit unique preferences for frequency signals, which provides us with a caching strategy to separate Attention and MLP to achieve an enhanced fit of exposure bias and reduce it. Based on this, we introduced FEB-Cache, a joint caching strategy that aligns with the non-exposed bias diffusion process (which gives us a higher performance cap) of caching Attention and MLP based on the frequency-guided cache table. Our approach combines a comprehensive understanding of the caching mechanism and offers a new perspective on leveraging caching to accelerate the diffusion process. Empirical results indicate that FEB-Cache optimizes model performance while concurrently facilitating acceleration.
Authors: Yuhang Ma, Bo Cheng, Shanyuan Liu, Hongyi Zhou, Liebucha Wu, Xiaoyu Wu, Dawei Leng, Yuhui Yin
Abstract: Flow-based Transformer models have achieved state-of-the-art image generation performance, but often suffer from high inference latency and computational cost due to their large parameter sizes. To improve inference efficiency without compromising quality, we propose Bridged Progressive Rectified Flow Transformers (NAMI), which decompose the generation process across temporal, spatial, and architectural demensions. We divide the rectified flow into different stages according to resolution, and use a BridgeFlow module to connect them. Fewer Transformer layers are used at low-resolution stages to generate image layouts and concept contours, and more layers are progressively added as the resolution increases. Experiments demonstrate that our approach achieves fast convergence and reduces inference time while ensuring generation quality. The main contributions of this paper are summarized as follows: (1) We introduce Bridged Progressive Rectified Flow Transformers that enable multi-resolution training, accelerating model convergence; (2) NAMI leverages piecewise flow and spatial cascading of Diffusion Transformer (DiT) to rapidly generate images, reducing inference time by 64% for generating 1024 resolution images; (3) We propose a BridgeFlow module to align flows between different stages; (4) We propose the NAMI-1K benchmark to evaluate human preference performance, aiming to mitigate distributional bias and comprehensively assess model effectiveness. The results show that our model is competitive with state-of-the-art models.
Authors: Boqian Li, Haiwen Feng, Zeyu Cai, Michael J. Black, Yuliang Xiu
Abstract: Fitting a body to a 3D clothed human point cloud is a common yet challenging task. Traditional optimization-based approaches use multi-stage pipelines that are sensitive to pose initialization, while recent learning-based methods often struggle with generalization across diverse poses and garment types. We propose Equivariant Tightness Fitting for Clothed Humans, or ETCH, a novel pipeline that estimates cloth-to-body surface mapping through locally approximate SE(3) equivariance, encoding tightness as displacement vectors from the cloth surface to the underlying body. Following this mapping, pose-invariant body features regress sparse body markers, simplifying clothed human fitting into an inner-body marker fitting task. Extensive experiments on CAPE and 4D-Dress show that ETCH significantly outperforms state-of-the-art methods -- both tightness-agnostic and tightness-aware -- in body fitting accuracy on loose clothing (16.7% ~ 69.5%) and shape accuracy (average 49.9%). Our equivariant tightness design can even reduce directional errors by (67.2% ~ 89.8%) in one-shot (or out-of-distribution) settings (~ 1% data). Qualitative results demonstrate strong generalization of ETCH, regardless of challenging poses, unseen shapes, loose clothing, and non-rigid dynamics. We will release the code and models soon for research purposes at https://boqian-li.github.io/ETCH/.
Authors: Qingyu Shi, Jianzong Wu, Jinbin Bai, Jiangning Zhang, Lu Qi, Yunhai Tong, Xiangtai Li
Abstract: The motion transfer task aims to transfer motion from a source video to newly generated videos, requiring the model to decouple motion from appearance. Previous diffusion-based methods primarily rely on separate spatial and temporal attention mechanisms within the 3D U-Net. In contrast, state-of-the-art video Diffusion Transformers (DiT) models use 3D full attention, which does not explicitly separate temporal and spatial information. Thus, the interaction between spatial and temporal dimensions makes decoupling motion and appearance more challenging for DiT models. In this paper, we propose DeT, a method that adapts DiT models to improve motion transfer ability. Our approach introduces a simple yet effective temporal kernel to smooth DiT features along the temporal dimension, facilitating the decoupling of foreground motion from background appearance. Meanwhile, the temporal kernel effectively captures temporal variations in DiT features, which are closely related to motion. Moreover, we introduce explicit supervision along dense trajectories in the latent feature space to further enhance motion consistency. Additionally, we present MTBench, a general and challenging benchmark for motion transfer. We also introduce a hybrid motion fidelity metric that considers both the global and local motion similarity. Therefore, our work provides a more comprehensive evaluation than previous works. Extensive experiments on MTBench demonstrate that DeT achieves the best trade-off between motion fidelity and edit fidelity.
Authors: Chi-Hsi Kung, Frangil Ramirez, Juhyung Ha, Yi-Ting Chen, David Crandall, Yi-Hsuan Tsai
Abstract: Understanding a procedural activity requires modeling both how action steps transform the scene, and how evolving scene transformations can influence the sequence of action steps, even those that are accidental or erroneous. Existing work has studied procedure-aware video representations by modeling the temporal order of actions, but has not explicitly learned the state changes (scene transformations). In this work, we study procedure-aware video representation learning by incorporating state-change descriptions generated by Large Language Models (LLMs) as supervision signals for video encoders. Moreover, we generate state-change counterfactuals that simulate hypothesized failure outcomes, allowing models to learn by imagining unseen "What if" scenarios. This counterfactual reasoning facilitates the model's ability to understand the cause and effect of each step in an activity. We conduct extensive experiments on procedure-aware tasks, including temporal action segmentation, error detection, action phase classification, frame retrieval, multi-instance retrieval, and action recognition. Our results demonstrate the effectiveness of the proposed state-change descriptions and their counterfactuals, and achieve significant improvements on multiple tasks. Code is available at https://github.com/HCIS- Lab/counterfactual-video-pretrain.
URLs: https://github.com/HCIS-
Authors: Nikai Du, Zhennan Chen, Shan Gao, Zhizhou Chen, Xi Chen, Zhengkai Jiang, Jian Yang, Ying Tai
Abstract: This paper explores the task of Complex Visual Text Generation (CVTG), which centers on generating intricate textual content distributed across diverse regions within visual images. In CVTG, image generation models often rendering distorted and blurred visual text or missing some visual text. To tackle these challenges, we propose TextCrafter, a novel multi-visual text rendering method. TextCrafter employs a progressive strategy to decompose complex visual text into distinct components while ensuring robust alignment between textual content and its visual carrier. Additionally, it incorporates a token focus enhancement mechanism to amplify the prominence of visual text during the generation process. TextCrafter effectively addresses key challenges in CVTG tasks, such as text confusion, omissions, and blurriness. Moreover, we present a new benchmark dataset, CVTG-2K, tailored to rigorously evaluate the performance of generative models on CVTG tasks. Extensive experiments demonstrate that our method surpasses state-of-the-art approaches.
Authors: Jannik Endres, Oliver Hahn, Charles Corbi\`ere, Simone Schaub-Meyer, Stefan Roth, Alexandre Alahi
Abstract: Omnidirectional depth perception is essential for mobile robotics applications that require scene understanding across a full 360{\deg} field of view. Camera-based setups offer a cost-effective option by using stereo depth estimation to generate dense, high-resolution depth maps without relying on expensive active sensing. However, existing omnidirectional stereo matching approaches achieve only limited depth accuracy across diverse environments, depth ranges, and lighting conditions, due to the scarcity of real-world data. We present DFI-OmniStereo, a novel omnidirectional stereo matching method that leverages a large-scale pre-trained foundation model for relative monocular depth estimation within an iterative optimization-based stereo matching architecture. We introduce a dedicated two-stage training strategy to utilize the relative monocular depth features for our omnidirectional stereo matching before scale-invariant fine-tuning. DFI-OmniStereo achieves state-of-the-art results on the real-world Helvipad dataset, reducing disparity MAE by approximately 16% compared to the previous best omnidirectional stereo method.
Authors: Tobias Fischer, Samuel Rota Bul\`o, Yung-Hsu Yang, Nikhil Keetha, Lorenzo Porzi, Norman M\"uller, Katja Schwarz, Jonathon Luiten, Marc Pollefeys, Peter Kontschieder
Abstract: 3D Gaussian splatting enables high-quality novel view synthesis (NVS) at real-time frame rates. However, its quality drops sharply as we depart from the training views. Thus, dense captures are needed to match the high-quality expectations of applications like Virtual Reality (VR). However, such dense captures are very laborious and expensive to obtain. Existing works have explored using 2D generative models to alleviate this requirement by distillation or generating additional training views. These models typically rely on a noise-to-data generative process conditioned only on a handful of reference input views, leading to hallucinations, inconsistent generation results, and subsequent reconstruction artifacts. Instead, we propose a multi-view, flow matching model that learns a flow to directly connect novel view renderings from possibly sparse reconstructions to renderings that we expect from dense reconstructions. This enables augmenting scene captures with consistent, generated views to improve reconstruction quality. Our model is trained on a novel dataset of 3.6M image pairs and can process up to 45 views at 540x960 resolution (91K tokens) on one H100 GPU in a single forward pass. Our pipeline consistently improves NVS in sparse- and dense-view scenarios, leading to higher-quality reconstructions than prior works across multiple, widely-used NVS benchmarks.
Authors: Kyungmin Choi, JaKeoung Koo, Stephen McLaughlin, Abderrahim Halimi
Abstract: Single-photon Lidar imaging offers a significant advantage in 3D imaging due to its high resolution and long-range capabilities, however it is challenging to apply in noisy environments with multiple targets per pixel. To tackle these challenges, several methods have been proposed. Statistical methods demonstrate interpretability on the inferred parameters, but they are often limited in their ability to handle complex scenes. Deep learning-based methods have shown superior performance in terms of accuracy and robustness, but they lack interpretability or they are limited to a single-peak per pixel. In this paper, we propose a deep unrolling algorithm for dual-peak single-photon Lidar imaging. We introduce a hierarchical Bayesian model for multiple targets and propose a neural network that unrolls the underlying statistical method. To support multiple targets, we adopt a dual depth maps representation and exploit geometric deep learning to extract features from the point cloud. The proposed method takes advantages of statistical methods and learning-based methods in terms of accuracy and quantifying uncertainty. The experimental results on synthetic and real data demonstrate the competitive performance when compared to existing methods, while also providing uncertainty information.
Authors: Sebastian Schmidt, Julius K\"orner, Dominik Fuchsgruber, Stefano Gasperini, Federico Tombari, Stephan G\"unnemann
Abstract: In panoptic segmentation, individual instances must be separated within semantic classes. As state-of-the-art methods rely on a pre-defined set of classes, they struggle with novel categories and out-of-distribution (OOD) data. This is particularly problematic in safety-critical applications, such as autonomous driving, where reliability in unseen scenarios is essential. We address the gap between outstanding benchmark performance and reliability by proposing Prior2Former (P2F), the first approach for segmentation vision transformers rooted in evidential learning. P2F extends the mask vision transformer architecture by incorporating a Beta prior for computing model uncertainty in pixel-wise binary mask assignments. This design enables high-quality uncertainty estimation that effectively detects novel and OOD objects enabling state-of-the-art anomaly instance segmentation and open-world panoptic segmentation. Unlike most segmentation models addressing unknown classes, P2F operates without access to OOD data samples or contrastive training on void (i.e., unlabeled) classes, making it highly applicable in real-world scenarios where such prior information is unavailable. Additionally, P2F can be flexibly applied to anomaly instance and panoptic segmentation. Through comprehensive experiments on the Cityscapes, COCO, SegmentMeIfYouCan, and OoDIS datasets, P2F demonstrates state-of-the-art performance across the board.
Authors: Zitian Tang, Shijie Wang, Junho Cho, Jaewook Yoo, Chen Sun
Abstract: Do we still need to represent objects explicitly in multimodal large language models (MLLMs)? To one extreme, pre-trained encoders convert images into visual tokens, with which objects and spatiotemporal relationships may be implicitly modeled. To the other extreme, image captions by themselves provide strong empirical performances for understanding tasks, despite missing fine-grained spatiotemporal information. To answer this question, we introduce ObjectMLLM, a framework capable of leveraging arbitrary computer vision algorithm to extract and integrate structured visual representation. Through extensive evaluations on six video question answering benchmarks, we confirm that explicit integration of object-centric representation remains necessary. Surprisingly, we observe that the simple approach of quantizing the continuous, structured object information and representing them as plain text performs the best, offering a data-efficient approach to integrate other visual perception modules into MLLM design. Our code and models are released at https://github.com/brown-palm/ObjectMLLM.
Authors: Jinghe Yang, Mingming Gong, Ye Pu
Abstract: Autonomous Underwater Vehicles (AUVs) play a crucial role in underwater exploration. Vision-based methods offer cost-effective solutions for localization and mapping in the absence of conventional sensors like GPS and LiDAR. However, underwater environments present significant challenges for feature extraction and matching due to image blurring and noise caused by attenuation, scattering, and the interference of \textit{marine snow}. In this paper, we aim to improve the robustness of the feature extraction and matching in the turbid underwater environment using the cross-modal knowledge distillation method that transfers the in-air feature extraction and matching models to underwater settings using synthetic underwater images as the medium. We first propose a novel adaptive GAN-synthesis method to estimate water parameters and underwater noise distribution, to generate environment-specific synthetic underwater images. We then introduce a general knowledge distillation framework compatible with different teacher models. The evaluation of GAN-based synthesis highlights the significance of the new components, i.e. GAN-synthesized noise and forward scattering, in the proposed model. Additionally, VSLAM, as a representative downstream application of feature extraction and matching, is employed on real underwater sequences to validate the effectiveness of the transferred model. Project page: https://github.com/Jinghe-mel/UFEN-GAN.
Authors: Jiawei Wu, Zhifei Yang, Zhe Wang, Zhi Jin
Abstract: All-in-one image restoration (AIR) aims to address diverse degradations within a unified model by leveraging informative degradation conditions to guide the restoration process. However, existing methods often rely on implicitly learned priors, which may entangle feature representations and hinder performance in complex or unseen scenarios. Histogram of Oriented Gradients (HOG) as a classical gradient representation, we observe that it has strong discriminative capability across diverse degradations, making it a powerful and interpretable prior for AIR. Based on this insight, we propose HOGformer, a Transformer-based model that integrates learnable HOG features for degradation-aware restoration. The core of HOGformer is a Dynamic HOG-aware Self-Attention (DHOGSA) mechanism, which adaptively models long-range spatial dependencies conditioned on degradation-specific cues encoded by HOG descriptors. To further adapt the heterogeneity of degradations in AIR, we propose a Dynamic Interaction Feed-Forward (DIFF) module that facilitates channel-spatial interactions, enabling robust feature transformation under diverse degradations. Besides, we propose a HOG loss to explicitly enhance structural fidelity and edge sharpness. Extensive experiments on a variety of benchmarks, including adverse weather and natural degradations, demonstrate that HOGformer achieves state-of-the-art performance and generalizes well to complex real-world scenarios.Code is available at https://github.com/Fire-friend/HOGformer.
Authors: Yan Shi, Qingdong He, Yijun Liu, Xiaoyu Liu, Jingyong Su
Abstract: Multi-Layer Perceptrons (MLPs) have become one of the fundamental architectural component in point cloud analysis due to its effective feature learning mechanism. However, when processing complex geometric structures in point clouds, MLPs' fixed activation functions struggle to efficiently capture local geometric features, while suffering from poor parameter efficiency and high model redundancy. In this paper, we propose PointKAN, which applies Kolmogorov-Arnold Networks (KANs) to point cloud analysis tasks to investigate their efficacy in hierarchical feature representation. First, we introduce a Geometric Affine Module (GAM) to transform local features, improving the model's robustness to geometric variations. Next, in the Local Feature Processing (LFP), a parallel structure extracts both group-level features and global context, providing a rich representation of both fine details and overall structure. Finally, these features are combined and processed in the Global Feature Processing (GFP). By repeating these operations, the receptive field gradually expands, enabling the model to capture complete geometric information of the point cloud. To overcome the high parameter counts and computational inefficiency of standard KANs, we develop Efficient-KANs in the PointKAN-elite variant, which significantly reduces parameters while maintaining accuracy. Experimental results demonstrate that PointKAN outperforms PointMLP on benchmark datasets such as ModelNet40, ScanObjectNN, and ShapeNetPart, with particularly strong performance in Few-shot Learning task. Additionally, PointKAN achieves substantial reductions in parameter counts and computational complexity (FLOPs). This work highlights the potential of KANs-based architectures in 3D vision and opens new avenues for research in point cloud understanding.
Authors: Tiancheng Gu, Kaicheng Yang, Ziyong Feng, Xingjun Wang, Yanzhao Zhang, Dingkun Long, Yingda Chen, Weidong Cai, Jiankang Deng
Abstract: The Contrastive Language-Image Pre-training (CLIP) framework has become a widely used approach for multimodal representation learning, particularly in image-text retrieval and clustering. However, its efficacy is constrained by three key limitations: (1) text token truncation, (2) isolated image-text encoding, and (3) deficient compositionality due to bag-of-words behavior. While recent Multimodal Large Language Models (MLLMs) have demonstrated significant advances in generalized vision-language understanding, their potential for learning transferable multimodal representations remains underexplored.In this work, we present UniME (Universal Multimodal Embedding), a novel two-stage framework that leverages MLLMs to learn discriminative representations for diverse downstream tasks. In the first stage, we perform textual discriminative knowledge distillation from a powerful LLM-based teacher model to enhance the embedding capability of the MLLM\'s language component. In the second stage, we introduce hard negative enhanced instruction tuning to further advance discriminative representation learning. Specifically, we initially mitigate false negative contamination and then sample multiple hard negatives per instance within each batch, forcing the model to focus on challenging samples. This approach not only improves discriminative power but also enhances instruction-following ability in downstream tasks. We conduct extensive experiments on the MMEB benchmark and multiple retrieval tasks, including short and long caption retrieval and compositional retrieval. Results demonstrate that UniME achieves consistent performance improvement across all tasks, exhibiting superior discriminative and compositional capabilities.
Authors: Woo Yi Yang, Jiarui Wang, Sijing Wu, Huiyu Duan, Yuxin Zhu, Liu Yang, Kang Fu, Guangtao Zhai, Xiongkuo Min
Abstract: The rapid advancement in generative artificial intelligence have enabled the creation of 3D human faces (HFs) for applications including media production, virtual reality, security, healthcare, and game development, etc. However, assessing the quality and realism of these AI-generated 3D human faces remains a significant challenge due to the subjective nature of human perception and innate perceptual sensitivity to facial features. To this end, we conduct a comprehensive study on the quality assessment of AI-generated 3D human faces. We first introduce Gen3DHF, a large-scale benchmark comprising 2,000 videos of AI-Generated 3D Human Faces along with 4,000 Mean Opinion Scores (MOS) collected across two dimensions, i.e., quality and authenticity, 2,000 distortion-aware saliency maps and distortion descriptions. Based on Gen3DHF, we propose LMME3DHF, a Large Multimodal Model (LMM)-based metric for Evaluating 3DHF capable of quality and authenticity score prediction, distortion-aware visual question answering, and distortion-aware saliency prediction. Experimental results show that LMME3DHF achieves state-of-the-art performance, surpassing existing methods in both accurately predicting quality scores for AI-generated 3D human faces and effectively identifying distortion-aware salient regions and distortion types, while maintaining strong alignment with human perceptual judgments. Both the Gen3DHF database and the LMME3DHF will be released upon the publication.
Authors: Kwon Byung-Ki, Qi Dai, Lee Hyoseok, Chong Luo, Tae-Hyun Oh
Abstract: We present JointDiT, a diffusion transformer that models the joint distribution of RGB and depth. By leveraging the architectural benefit and outstanding image prior of the state-of-the-art diffusion transformer, JointDiT not only generates high-fidelity images but also produces geometrically plausible and accurate depth maps. This solid joint distribution modeling is achieved through two simple yet effective techniques that we propose, namely, adaptive scheduling weights, which depend on the noise levels of each modality, and the unbalanced timestep sampling strategy. With these techniques, we train our model across all noise levels for each modality, enabling JointDiT to naturally handle various combinatorial generation tasks, including joint generation, depth estimation, and depth-conditioned image generation by simply controlling the timesteps of each branch. JointDiT demonstrates outstanding joint generation performance. Furthermore, it achieves comparable results in depth estimation and depth-conditioned image generation, suggesting that joint distribution modeling can serve as a viable alternative to conditional generation. The project page is available at https://byungki-k.github.io/JointDiT/.
Authors: Rui Lan, Yancheng Bai, Xu Duan, Mingxing Li, Dongyang Jin, Ryan Xu, Lei Sun, Xiangxiang Chu
Abstract: Scene text editing aims to modify or add texts on images while ensuring text fidelity and overall visual quality consistent with the background. Recent methods are primarily built on UNet-based diffusion models, which have improved scene text editing results, but still struggle with complex glyph structures, especially for non-Latin ones (\eg, Chinese, Korean, Japanese). To address these issues, we present \textbf{FLUX-Text}, a simple and advanced multilingual scene text editing DiT method. Specifically, our FLUX-Text enhances glyph understanding and generation through lightweight Visual and Text Embedding Modules, while preserving the original generative capability of FLUX. We further propose a Regional Text Perceptual Loss tailored for text regions, along with a matching two-stage training strategy to better balance text editing and overall image quality. Benefiting from the DiT-based architecture and lightweight feature injection modules, FLUX-Text can be trained with only $0.1$M training examples, a \textbf{97\%} reduction compared to $2.9$M required by popular methods. Extensive experiments on multiple public datasets, including English and Chinese benchmarks, demonstrate that our method surpasses other methods in visual quality and text fidelity. All the code is available at https://github.com/AMAP-ML/FluxText.
Authors: Ciyu Ruan, Ruishan Guo, Zihang Gong, Jingao Xu, Wenhan Yang, Xinlei Chen
Abstract: Event cameras excel in high temporal resolution and dynamic range but suffer from dense noise in rainy conditions. Existing event deraining methods face trade-offs between temporal precision, deraining effectiveness, and computational efficiency. In this paper, we propose PRE-Mamba, a novel point-based event camera deraining framework that fully exploits the spatiotemporal characteristics of raw event and rain. Our framework introduces a 4D event cloud representation that integrates dual temporal scales to preserve high temporal precision, a Spatio-Temporal Decoupling and Fusion module (STDF) that enhances deraining capability by enabling shallow decoupling and interaction of temporal and spatial information, and a Multi-Scale State Space Model (MS3M) that captures deeper rain dynamics across dual-temporal and multi-spatial scales with linear computational complexity. Enhanced by frequency-domain regularization, PRE-Mamba achieves superior performance (0.95 SR, 0.91 NR, and 0.4s/M events) with only 0.26M parameters on EventRain-27K, a comprehensive dataset with labeled synthetic and real-world sequences. Moreover, our method generalizes well across varying rain intensities, viewpoints, and even snowy conditions.
Authors: Benjamin Raphael Ernhofer, Daniil Prokhorov, Jannica Langner, Dominik Bollmann
Abstract: Modern automotive infotainment systems necessitate intelligent and adaptive solutions to manage frequent User Interface (UI) updates and diverse design variations. This work introduces a vision-language framework to facilitate the understanding of and interaction with automotive UIs, enabling seamless adaptation across different UI designs. To support research in this field, AutomotiveUI-Bench-4K, an open-source dataset comprising 998 images with 4,208 annotations, is also released. Additionally, a data pipeline for generating training data is presented. A Molmo-7B-based model is fine-tuned using Low-Rank Adaptation (LoRa), incorporating generated reasoning along with visual grounding and evaluation capabilities. The fine-tuned Evaluative Large Action Model (ELAM) achieves strong performance on AutomotiveUI-Bench-4K (model and dataset are available on Hugging Face). The approach demonstrates strong cross-domain generalization, including a +5.6% improvement on ScreenSpot over the baseline model. An average accuracy of 80.8% is achieved on ScreenSpot, closely matching or surpassing specialized models for desktop, mobile, and web, despite being trained primarily on the automotive domain. This research investigates how data collection and subsequent fine-tuning can lead to AI-driven advancements in automotive UI understanding and interaction. The applied method is cost-efficient, and fine-tuned models can be deployed on consumer-grade GPUs.
Authors: Qingmei Li, Yang Zhang, Zurong Mai, Yuhang Chen, Shuohong Lou, Henglian Huang, Jiarui Zhang, Zhiwei Zhang, Yibin Wen, Weijia Li, Haohuan Fu, Jianxi Huang, Juepeng Zheng
Abstract: Large Multimodal Models (LMMs) has demonstrated capabilities across various domains, but comprehensive benchmarks for agricultural remote sensing (RS) remain scarce. Existing benchmarks designed for agricultural RS scenarios exhibit notable limitations, primarily in terms of insufficient scene diversity in the dataset and oversimplified task design. To bridge this gap, we introduce AgroMind, a comprehensive agricultural remote sensing benchmark covering four task dimensions: spatial perception, object understanding, scene understanding, and scene reasoning, with a total of 13 task types, ranging from crop identification and health monitoring to environmental analysis. We curate a high-quality evaluation set by integrating eight public datasets and one private farmland plot dataset, containing 27,247 QA pairs and 19,615 images. The pipeline begins with multi-source data pre-processing, including collection, format standardization, and annotation refinement. We then generate a diverse set of agriculturally relevant questions through the systematic definition of tasks. Finally, we employ LMMs for inference, generating responses, and performing detailed examinations. We evaluated 20 open-source LMMs and 4 closed-source models on AgroMind. Experiments reveal significant performance gaps, particularly in spatial reasoning and fine-grained recognition, it is notable that human performance lags behind several leading LMMs. By establishing a standardized evaluation framework for agricultural RS, AgroMind reveals the limitations of LMMs in domain knowledge and highlights critical challenges for future work. Data and code can be accessed at https://rssysu.github.io/AgroMind/.
Authors: Yuxuan Wang, Xuanyu Yi, Qingshan Xu, Yuan Zhou, Long Chen, Hanwang Zhang
Abstract: Personalizing 3D scenes from a single reference image enables intuitive user-guided editing, which requires achieving both multi-view consistency across perspectives and referential consistency with the input image. However, these goals are particularly challenging due to the viewpoint bias caused by the limited perspective provided in a single image. Lacking the mechanisms to effectively expand reference information beyond the original view, existing methods of image-conditioned 3DGS personalization often suffer from this viewpoint bias and struggle to produce consistent results. Therefore, in this paper, we present Consistent Personalization for 3D Gaussian Splatting (CP-GS), a framework that progressively propagates the single-view reference appearance to novel perspectives. In particular, CP-GS integrates pre-trained image-to-3D generation and iterative LoRA fine-tuning to extract and extend the reference appearance, and finally produces faithful multi-view guidance images and the personalized 3DGS outputs through a view-consistent generation process guided by geometric cues. Extensive experiments on real-world scenes show that our CP-GS effectively mitigates the viewpoint bias, achieving high-quality personalization that significantly outperforms existing methods. The code will be released at https://github.com/Yuxuan-W/CP-GS.
Authors: Xiaorong Zhu, Ziheng Jia, Jiarui Wang, Xiangyu Zhao, Haodong Duan, Xiongkuo Min, Jia Wang, Zicheng Zhang, Guangtao Zhai
Abstract: The rapid evolution of Multi-modality Large Language Models (MLLMs) is driving significant advancements in visual understanding and generation. Nevertheless, a comprehensive assessment of their capabilities, concerning the fine-grained physical principles especially in geometric optics, remains underexplored. To address this gap, we introduce GOBench, the first benchmark to systematically evaluate MLLMs' ability across two tasks: 1) Generating Optically Authentic Imagery and 2) Understanding Underlying Optical Phenomena. We curates high-quality prompts of geometric optical scenarios and use MLLMs to construct GOBench-Gen-1k dataset.We then organize subjective experiments to assess the generated imagery based on Optical Authenticity, Aesthetic Quality, and Instruction Fidelity, revealing MLLMs' generation flaws that violate optical principles. For the understanding task, we apply crafted evaluation instructions to test optical understanding ability of eleven prominent MLLMs. The experimental results demonstrate that current models face significant challenges in both optical generation and understanding. The top-performing generative model, GPT-4o-Image, cannot perfectly complete all generation tasks, and the best-performing MLLM model, Gemini-2.5Pro, attains a mere 37.35\% accuracy in optical understanding. Database and codes are publicly available at https://github.com/aiben-ch/GOBench.
Authors: Jinting Wang, Shan Yang, Chenxing Li, Dong Yu, Li Liu
Abstract: Cued Speech (CS) enhances lipreading via hand coding, offering visual phonemic cues that support precise speech perception for the hearing-impaired. The task of CS Video-to-Speech generation (CSV2S) aims to convert CS videos into intelligible speech signals. Most existing research focuses on CS Recognition (CSR), which transcribes video content into text. Consequently, a common solution for CSV2S is to integrate CSR with a text-to-speech (TTS) system. However, this pipeline relies on text as an intermediate medium, which may lead to error propagation and temporal misalignment between speech and CS video dynamics. In contrast, directly generating audio speech from CS video (direct CSV2S) often suffers from the inherent multimodal complexity and the limited availability of CS data. To address these challenges, we propose UniCUE, the first unified framework for CSV2S that directly generates speech from CS videos without relying on intermediate text. The core innovation of UniCUE lies in integrating an understanding task (CSR) that provides fine-grained CS visual-semantic cues to guide speech generation. Specifically, UniCUE incorporates a pose-aware visual processor, a semantic alignment pool that enables precise visual-semantic mapping, and a VisioPhonetic adapter to bridge the understanding and generation tasks within a unified architecture. To support this framework, we construct UniCUE-HI, a large-scale Mandarin CS dataset containing 11282 videos from 14 cuers, including both hearing-impaired and normal-hearing individuals. Extensive experiments on this dataset demonstrate that UniCUE achieves state-of-the-art performance across multiple evaluation metrics.
Authors: Nan Wang, Yuantao Chen, Lixing Xiao, Weiqing Xiao, Bohan Li, Zhaoxi Chen, Chongjie Ye, Shaocong Xu, Saining Zhang, Ziyang Yan, Pierre Merriaux, Lei Lei, Tianfan Xue, Hao Zhao
Abstract: Neural rendering techniques, including NeRF and Gaussian Splatting (GS), rely on photometric consistency to produce high-quality reconstructions. However, in real-world scenarios, it is challenging to guarantee perfect photometric consistency in acquired images. Appearance codes have been widely used to address this issue, but their modeling capability is limited, as a single code is applied to the entire image. Recently, the bilateral grid was introduced to perform pixel-wise color mapping, but it is difficult to optimize and constrain effectively. In this paper, we propose a novel multi-scale bilateral grid that unifies appearance codes and bilateral grids. We demonstrate that this approach significantly improves geometric accuracy in dynamic, decoupled autonomous driving scene reconstruction, outperforming both appearance codes and bilateral grids. This is crucial for autonomous driving, where accurate geometry is important for obstacle avoidance and control. Our method shows strong results across four datasets: Waymo, NuScenes, Argoverse, and PandaSet. We further demonstrate that the improvement in geometry is driven by the multi-scale bilateral grid, which effectively reduces floaters caused by photometric inconsistency.
Authors: Ivan Rodin, Tz-Ying Wu, Kyle Min, Sharath Nittur Sridhar, Antonino Furnari, Subarna Tripathi, Giovanni Maria Farinella
Abstract: We introduce EASG-Bench, a question-answering benchmark for egocentric videos where the question-answering pairs are created from spatio-temporally grounded dynamic scene graphs capturing intricate relationships among actors, actions, and objects. We propose a systematic evaluation framework and evaluate several language-only and video large language models (video-LLMs) on this benchmark. We observe a performance gap in language-only and video-LLMs, especially on questions focusing on temporal ordering, thus identifying a research gap in the area of long-context video understanding. To promote the reproducibility of our findings and facilitate further research, the benchmark and accompanying code are available at the following GitHub page: https://github.com/fpv-iplab/EASG-bench.
Authors: Wei Zhang, Yuantao Wang, Haowei Yang, Yin Zhuang, Shijian Lu, Xuerui Mao
Abstract: Real-world object detection is a challenging task where the captured images/videos often suffer from complex degradations due to various adverse weather conditions such as rain, fog, snow, low-light, etc. Despite extensive prior efforts, most existing methods are designed for one specific type of adverse weather with constraints of poor generalization, under-utilization of visual features while handling various image degradations. Leveraging a theoretical analysis on how critical visual details are lost in adverse-weather images, we design UniDet-D, a unified framework that tackles the challenge of object detection under various adverse weather conditions, and achieves object detection and image restoration within a single network. Specifically, the proposed UniDet-D incorporates a dynamic spectral attention mechanism that adaptively emphasizes informative spectral components while suppressing irrelevant ones, enabling more robust and discriminative feature representation across various degradation types. Extensive experiments show that UniDet-D achieves superior detection accuracy across different types of adverse-weather degradation. Furthermore, UniDet-D demonstrates superior generalization towards unseen adverse weather conditions such as sandstorms and rain-fog mixtures, highlighting its great potential for real-world deployment.
Authors: Youze Wang, Zijun Chen, Ruoyu Chen, Shishen Gu, Wenbo Hu, Jiayang Liu, Yinpeng Dong, Hang Su, Jun Zhu, Meng Wang, Richang Hong
Abstract: Recent advancements in multimodal large language models for video understanding (videoLLMs) have enhanced their capacity to process complex spatiotemporal data. However, challenges such as factual inaccuracies, harmful content, biases, hallucinations, and privacy risks compromise their reliability. This study introduces Trust-videoLLMs, a first comprehensive benchmark evaluating 23 state-of-the-art videoLLMs (5 commercial, 18 open-source) across five critical dimensions: truthfulness, robustness, safety, fairness, and privacy. Comprising 30 tasks with adapted, synthetic, and annotated videos, the framework assesses spatiotemporal risks, temporal consistency and cross-modal impact. Results reveal significant limitations in dynamic scene comprehension, cross-modal perturbation resilience and real-world risk mitigation. While open-source models occasionally outperform, proprietary models generally exhibit superior credibility, though scaling does not consistently improve performance. These findings underscore the need for enhanced training datat diversity and robust multimodal alignment. Trust-videoLLMs provides a publicly available, extensible toolkit for standardized trustworthiness assessments, addressing the critical gap between accuracy-focused benchmarks and demands for robustness, safety, fairness, and privacy.
Authors: Nikos Spyrou, Athanasios Vlontzos, Paraskevas Pegios, Thomas Melistas, Nefeli Gkouti, Yannis Panagakis, Giorgos Papanastasiou, Sotirios A. Tsaftaris
Abstract: Adapting text-to-image (T2I) latent diffusion models (LDMs) to video editing has shown strong visual fidelity and controllability, but challenges remain in maintaining causal relationships inherent to the video data generating process. Edits affecting causally dependent attributes often generate unrealistic or misleading outcomes if these relationships are ignored. In this work, we introduce a causally faithful framework for counterfactual video generation, formulated as an Out-of-Distribution (OOD) prediction problem. We embed prior causal knowledge by encoding the relationships specified in a causal graph into text prompts and guide the generation process by optimizing these prompts using a vision-language model (VLM)-based textual loss. This loss encourages the latent space of the LDMs to capture OOD variations in the form of counterfactuals, effectively steering generation toward causally meaningful alternatives. The proposed framework, dubbed CSVC, is agnostic to the underlying video editing system and does not require access to its internal mechanisms or fine-tuning. We evaluate our approach using standard video quality metrics and counterfactual-specific criteria, such as causal effectiveness and minimality. Experimental results show that CSVC generates causally faithful video counterfactuals within the LDM distribution via prompt-based causal steering, achieving state-of-the-art causal effectiveness without compromising temporal consistency or visual quality on real-world facial videos. Due to its compatibility with any black-box video editing system, our framework has significant potential to generate realistic 'what if' hypothetical video scenarios in diverse areas such as digital media and healthcare.
Authors: Xiuyu Yang, Shuhan Tan, Philipp Kr\"ahenb\"uhl
Abstract: An ideal traffic simulator replicates the realistic long-term point-to-point trip that a self-driving system experiences during deployment. Prior models and benchmarks focus on closed-loop motion simulation for initial agents in a scene. This is problematic for long-term simulation. Agents enter and exit the scene as the ego vehicle enters new regions. We propose InfGen, a unified next-token prediction model that performs interleaved closed-loop motion simulation and scene generation. InfGen automatically switches between closed-loop motion simulation and scene generation mode. It enables stable long-term rollout simulation. InfGen performs at the state-of-the-art in short-term (9s) traffic simulation, and significantly outperforms all other methods in long-term (30s) simulation. The code and model of InfGen will be released at https://orangesodahub.github.io/InfGen
Authors: Tongshun Zhang, Pingping Liu, Mengen Cai, Zijian Zhang, Yubing Lu, Qiuzhan Zhou
Abstract: Current low-light image enhancement (LLIE) methods face significant limitations in simultaneously improving brightness while preserving semantic consistency, fine details, and computational efficiency. With the emergence of state-space models, particularly Mamba, image restoration has achieved remarkable performance, yet existing visual Mamba approaches flatten 2D images into 1D token sequences using fixed scanning rules, critically limiting interactions between distant tokens with causal relationships and constraining their ability to capture meaningful long-range dependencies. To address these fundamental limitations, we propose BSMamba, a novel visual Mamba architecture comprising two specially designed components: Brightness Mamba and Semantic Mamba. The Brightness Mamba revolutionizes token interaction patterns by prioritizing connections between distant tokens with similar brightness levels, effectively addressing the challenge of brightness restoration in LLIE tasks through brightness-guided selective attention. Complementing this, the Semantic Mamba establishes priority interactions between tokens sharing similar semantic meanings, allowing the model to maintain contextual consistency by connecting semantically related regions across the image, thus preserving the hierarchical nature of image semantics during enhancement. By intelligently modeling tokens based on brightness and semantic similarity rather than arbitrary scanning patterns, BSMamba transcends the constraints of conventional token sequencing while adhering to the principles of causal modeling. Extensive experiments demonstrate that BSMamba achieves state-of-the-art performance in LLIE while preserving semantic consistency. Code is available at https://github.com/bywlzts/BSMamba.
Authors: Zunhui Xia, Hongxing Li, Libin Lan
Abstract: Medical image recognition serves as a key way to aid in clinical diagnosis, enabling more accurate and timely identification of diseases and abnormalities. Vision transformer-based approaches have proven effective in handling various medical recognition tasks. However, these methods encounter two primary challenges. First, they are often task-specific and architecture-tailored, limiting their general applicability. Second, they usually either adopt full attention to model long-range dependencies, resulting in high computational costs, or rely on handcrafted sparse attention, potentially leading to suboptimal performance. To tackle these issues, we present MedFormer, an efficient medical vision transformer with two key ideas. First, it employs a pyramid scaling structure as a versatile backbone for various medical image recognition tasks, including image classification and dense prediction tasks such as semantic segmentation and lesion detection. This structure facilitates hierarchical feature representation while reducing the computation load of feature maps, highly beneficial for boosting performance. Second, it introduces a novel Dual Sparse Selection Attention (DSSA) with content awareness to improve computational efficiency and robustness against noise while maintaining high performance. As the core building technique of MedFormer, DSSA is designed to explicitly attend to the most relevant content. Theoretical analysis demonstrates that MedFormer outperforms existing medical vision transformers in terms of generality and efficiency. Extensive experiments across various imaging modality datasets show that MedFormer consistently enhances performance in all three medical image recognition tasks mentioned above. MedFormer provides an efficient and versatile solution for medical image recognition, with strong potential for clinical application.
Authors: Ziying Song, Lin Liu, Hongyu Pan, Bencheng Liao, Mingzhe Guo, Lei Yang, Yongchang Zhang, Shaoqing Xu, Caiyan Jia, Yadan Luo
Abstract: Most end-to-end autonomous driving methods rely on imitation learning from single expert demonstrations, often leading to conservative and homogeneous behaviors that limit generalization in complex real-world scenarios. In this work, we propose DIVER, an end-to-end driving framework that integrates reinforcement learning with diffusion-based generation to produce diverse and feasible trajectories. At the core of DIVER lies a reinforced diffusion-based generation mechanism. First, the model conditions on map elements and surrounding agents to generate multiple reference trajectories from a single ground-truth trajectory, alleviating the limitations of imitation learning that arise from relying solely on single expert demonstrations. Second, reinforcement learning is employed to guide the diffusion process, where reward-based supervision enforces safety and diversity constraints on the generated trajectories, thereby enhancing their practicality and generalization capability. Furthermore, to address the limitations of L2-based open-loop metrics in capturing trajectory diversity, we propose a novel Diversity metric to evaluate the diversity of multi-mode predictions.Extensive experiments on the closed-loop NAVSIM and Bench2Drive benchmarks, as well as the open-loop nuScenes dataset, demonstrate that DIVER significantly improves trajectory diversity, effectively addressing the mode collapse problem inherent in imitation learning.
Authors: Zhitao Wang, Hengyu Man, Wenrui Li, Xingtao Wang, Xiaopeng Fan, Debin Zhao
Abstract: Recent advances in video generation techniques have given rise to an emerging paradigm of generative video coding for Ultra-Low Bitrate (ULB) scenarios by leveraging powerful generative priors. However, most existing methods are limited by domain specificity (e.g., facial or human videos) or excessive dependence on high-level text guidance, which tend to inadequately capture fine-grained motion details, leading to unrealistic or incoherent reconstructions. To address these challenges, we propose Trajectory-Guided Generative Video Coding (dubbed T-GVC), a novel framework that bridges low-level motion tracking with high-level semantic understanding. T-GVC features a semantic-aware sparse motion sampling pipeline that extracts pixel-wise motion as sparse trajectory points based on their semantic importance, significantly reducing the bitrate while preserving critical temporal semantic information. In addition, by integrating trajectory-aligned loss constraints into diffusion processes, we introduce a training-free guidance mechanism in latent space to ensure physically plausible motion patterns without sacrificing the inherent capabilities of generative models. Experimental results demonstrate that T-GVC outperforms both traditional and neural video codecs under ULB conditions. Furthermore, additional experiments confirm that our framework achieves more precise motion control than existing text-guided methods, paving the way for a novel direction of generative video coding guided by geometric motion modeling.
Authors: Junxue Yang, Xin Liao, Weixuan Tang, Jianhua Yang, Zheng Qin
Abstract: Image hiding fully explores the hidden potential of deep learning-based models, aiming to conceal image-level messages within cover images and reveal them from stego images to achieve covert communication. Existing hiding schemes are easily detected by the naked eyes or steganalyzers due to the cover type confined to the spatial domain, single-range feature extraction and attacks, and insufficient loss constraints. To address these issues, we propose a multi-range representations-driven adversarial stego generation framework called MRAG for JPEG image hiding. This design stems from the fact that steganalyzers typically combine local-range and global-range information to better capture hidden traces. Specifically, MRAG integrates the local-range characteristic of the convolution and the global-range modeling of the transformer. Meanwhile, a features angle-norm disentanglement loss is designed to launch multi-range representations-driven feature-level adversarial attacks. It computes the adversarial loss between covers and stegos based on the surrogate steganalyzer's classified features, i.e., the features before the last fully connected layer. Under the dual constraints of features angle and norm, MRAG can delicately encode the concatenation of cover and secret into subtle adversarial perturbations from local and global ranges relevant to steganalysis. Therefore, the resulting stego can achieve visual and steganalysis imperceptibility. Moreover, coarse-grained and fine-grained frequency decomposition operations are devised to transform the input, introducing multi-grained information. Extensive experiments demonstrate that MRAG can achieve state-of-the-art performance.
Authors: Jihao Gu, Fei Wang, Kun Li, Yanyan Wei, Zhiliang Wu, Dan Guo
Abstract: In this paper, we present MM-Gesture, the solution developed by our team HFUT-VUT, which ranked 1st in the micro-gesture classification track of the 3rd MiGA Challenge at IJCAI 2025, achieving superior performance compared to previous state-of-the-art methods. MM-Gesture is a multimodal fusion framework designed specifically for recognizing subtle and short-duration micro-gestures (MGs), integrating complementary cues from joint, limb, RGB video, Taylor-series video, optical-flow video, and depth video modalities. Utilizing PoseConv3D and Video Swin Transformer architectures with a novel modality-weighted ensemble strategy, our method further enhances RGB modality performance through transfer learning pre-trained on the larger MA-52 dataset. Extensive experiments on the iMiGUE benchmark, including ablation studies across different modalities, validate the effectiveness of our proposed approach, achieving a top-1 accuracy of 73.213%. Code is available at: https://github.com/momiji-bit/MM-Gesture.
Authors: Jaehyun Lee, Wonhark Park, Wonsik Shin, Hyunho Lee, Hyoung Min Na, Nojun Kwak
Abstract: Diffusion-based text-to-image models have achieved remarkable results in synthesizing diverse images from text prompts and can capture specific artistic styles via style personalization. However, their entangled latent space and lack of smooth interpolation make it difficult to apply distinct painting techniques in a controlled, regional manner, often causing one style to dominate. To overcome this, we propose a zero-shot diffusion pipeline that naturally blends multiple styles by performing style composition on the denoised latents predicted during the flow-matching denoising process of separately trained, style-specialized models. We leverage the fact that lower-noise latents carry stronger stylistic information and fuse them across heterogeneous diffusion pipelines using spatial masks, enabling precise, region-specific style control. This mechanism preserves the fidelity of each individual style while allowing user-guided mixing. Furthermore, to ensure structural coherence across different models, we incorporate depth-map conditioning via ControlNet into the diffusion framework. Qualitative and quantitative experiments demonstrate that our method successfully achieves region-specific style mixing according to the given masks.
Authors: Jiawei Xu, Kai Deng, Zexin Fan, Shenlong Wang, Jin Xie, Jian Yang
Abstract: Modeling and rendering dynamic urban driving scenes is crucial for self-driving simulation. Current high-quality methods typically rely on costly manual object tracklet annotations, while self-supervised approaches fail to capture dynamic object motions accurately and decompose scenes properly, resulting in rendering artifacts. We introduce AD-GS, a novel self-supervised framework for high-quality free-viewpoint rendering of driving scenes from a single log. At its core is a novel learnable motion model that integrates locality-aware B-spline curves with global-aware trigonometric functions, enabling flexible yet precise dynamic object modeling. Rather than requiring comprehensive semantic labeling, AD-GS automatically segments scenes into objects and background with the simplified pseudo 2D segmentation, representing objects using dynamic Gaussians and bidirectional temporal visibility masks. Further, our model incorporates visibility reasoning and physically rigid regularization to enhance robustness. Extensive evaluations demonstrate that our annotation-free model significantly outperforms current state-of-the-art annotation-free methods and is competitive with annotation-dependent approaches.
Authors: Junjie Hu, Tianyang Han, Kai Ma, Jialin Gao, Hao Dou, Song Yang, Xianhua He, Jianhui Zhang, Junfeng Luo, Xiaoming Wei, Wenqiang Zhang
Abstract: Recent subject-driven image customization has achieved significant advancements in fidelity, yet fine-grained instance-level spatial control remains elusive, hindering broader real-world application. This limitation is mainly attributed to the absence of scalable datasets that bind identity with precise positional cues. To this end, we introduce PositionIC, a unified framework that enforces position and identity consistency for multi-subject customization. We construct a scalable synthesis pipeline that employs a bidirectional generation paradigm to eliminate subject drift and maintain semantic coherence. On top of these data, we design a lightweight positional modulation operation that decouples spatial embeddings among subjects, enabling independent, accurate placement while preserving visual fidelity. Extensive experiments demonstrate that our approach can achieve precise spatial control while maintaining high consistency in image customization tasks. PositionIC paves the way for controllable, high-fidelity image customization in open-world, multi-entity scenarios and will be released to foster further research.
Authors: Shashanka Venkataramanan, Valentinos Pariza, Mohammadreza Salehi, Lukas Knobel, Spyros Gidaris, Elias Ramzi, Andrei Bursuc, Yuki M. Asano
Abstract: We present Franca (pronounced Fran-ka): free one; the first fully open-source (data, code, weights) vision foundation model that matches and in many cases surpasses the performance of state-of-the-art proprietary models, e.g., DINOv2, CLIP, SigLIPv2, etc. Our approach is grounded in a transparent training pipeline inspired by Web-SSL and uses publicly available data: ImageNet-21K and a subset of ReLAION-2B. Beyond model release, we tackle critical limitations in SSL clustering methods. While modern models rely on assigning image features to large codebooks via clustering algorithms like Sinkhorn-Knopp, they fail to account for the inherent ambiguity in clustering semantics. To address this, we introduce a parameter-efficient, multi-head clustering projector based on nested Matryoshka representations. This design progressively refines features into increasingly fine-grained clusters without increasing the model size, enabling both performance and memory efficiency. Additionally, we propose a novel positional disentanglement strategy that explicitly removes positional biases from dense representations, thereby improving the encoding of semantic content. This leads to consistent gains on several downstream benchmarks, demonstrating the utility of cleaner feature spaces. Our contributions establish a new standard for transparent, high-performance vision models and open a path toward more reproducible and generalizable foundation models for the broader AI community. The code and model checkpoints are available at https://github.com/valeoai/Franca.
Authors: Goeric Huybrechts, Srikanth Ronanki, Sai Muralidhar Jayanthi, Jack Fitzgerald, Srinivasan Veeravanallur
Abstract: The proliferation of multimodal Large Language Models has significantly advanced the ability to analyze and understand complex data inputs from different modalities. However, the processing of long documents remains under-explored, largely due to a lack of suitable benchmarks. To address this, we introduce Document Haystack, a comprehensive benchmark designed to evaluate the performance of Vision Language Models (VLMs) on long, visually complex documents. Document Haystack features documents ranging from 5 to 200 pages and strategically inserts pure text or multimodal text+image "needles" at various depths within the documents to challenge VLMs' retrieval capabilities. Comprising 400 document variants and a total of 8,250 questions, it is supported by an objective, automated evaluation framework. We detail the construction and characteristics of the Document Haystack dataset, present results from prominent VLMs and discuss potential research avenues in this area.
Authors: Minje Park, Jeonghwa Lim, Taehyung Yu, Sunghoon Joo
Abstract: Electrocardiogram (ECG) delineation, the segmentation of meaningful waveform features, is critical for clinical diagnosis. Despite recent advances using deep learning, progress has been limited by the scarcity of publicly available annotated datasets. Semi-supervised learning presents a promising solution by leveraging abundant unlabeled ECG data. In this study, we present SemiSegECG, the first systematic benchmark for semi-supervised semantic segmentation (SemiSeg) in ECG delineation. We curated and unified multiple public datasets, including previously underused sources, to support robust and diverse evaluation. We adopted five representative SemiSeg algorithms from computer vision, implemented them on two different architectures: the convolutional network and the transformer, and evaluated them in two different settings: in-domain and cross-domain. Additionally, we propose ECG-specific training configurations and augmentation strategies and introduce a standardized evaluation framework. Our results show that the transformer outperforms the convolutional network in semi-supervised ECG delineation. We anticipate that SemiSegECG will serve as a foundation for advancing semi-supervised ECG delineation methods and will facilitate further research in this domain.
Authors: Haochen Han, Alex Jinpeng Wang, Fangming Liu, Jun Zhu
Abstract: In this paper, we study a practical but less-touched problem in Vision-Language Models (VLMs), \ie, negation understanding. Specifically, many real-world applications require models to explicitly identify what is false or non-existent, \eg, radiologists may search for images that exclude specific conditions. Despite the impressive transferability of VLMs through large-scale training, they suffer from a critical limitation that fails to handle negation. To address this challenge, existing methods attribute its root cause to the scarcity of negation training data and propose to fine-tune VLMs on massive data containing explicit negation. Undoubtedly, such data-centric solutions demand substantial data and computational resources, limiting their sustainable widespread adoption. To tackle negation in a low-carbon manner, we empirically observe that the key obstacle lies in the dual-concept shifts between the affirmation and negation distributions. Therefore, we propose a Negation-Aware Test-Time Adaptation (NEAT) method to efficiently adjust distribution-related parameters during inference. In brief, NEAT can reduce distribution shift in consistent semantics while eliminating false distributional consistency in unrelated semantics. Extensive experiments on the various negation understanding tasks verify the effectiveness of the proposed method. Remarkably, with less than 0.01\% of trainable parameters, NEAT achieves comparable or superior performance to state-of-the-art post-training approaches. Our code is available at https://github.com/hhc1997/NEAT.
Authors: Yang Chen, Yufan Shen, Wenxuan Huang, Sheng Zhou, Qunshu Lin, Xinyu Cai, Zhi Yu, Jiajun Bu, Botian Shi, Yu Qiao
Abstract: Multimodal Large Language Models (MLLMs) exhibit impressive performance across various visual tasks. Subsequent investigations into enhancing their visual reasoning abilities have significantly expanded their performance envelope. However, a critical bottleneck in the advancement of MLLMs toward deep visual reasoning is their heavy reliance on curated image-text supervision. To solve this problem, we introduce a novel framework, ``Reasoning-Rendering-Visual-Feedback'' (RRVF), that enables MLLMs to learn complex visual reasoning from only raw images. This framework builds on the ``Asymmetry of Verification'' principle, i.e., verifying the rendered output against the source image is substantially easier than performing deep visual reasoning to generate a faithful, structured representation such as code. We demonstrate that this relative ease provides an ideal reward signal for optimization via Reinforcement Learning (RL), thereby reducing reliance on image-text supervision. RRVF implements a closed-loop iterative process encompassing reasoning, rendering, and visual feedback components, enabling the model to perform complex reasoning, including self-correction through multi-turn interactions. This process is optimized end-to-end using the GRPO algorithm. Extensive evaluations are conducted on image-to-code generation across two diverse domains: data charts and web interfaces. The RRVF-trained model not only outperforms existing similarly sized open-source MLLMs and supervised fine-tuning baselines but also exhibits superior generalization. Notably, the model outperforms the more advanced MLLM used to generate visual feedback during training. %only in arxiv Code is available at https://github.com/L-O-I/RRVF.
Authors: Sheng-Feng Yu, Jia-Jiun Yao, Wei-Chen Chiu
Abstract: Although larger datasets are crucial for training large deep models, the rapid growth of dataset size has brought a significant challenge in terms of considerable training costs, which even results in prohibitive computational expenses. Dataset Distillation becomes a popular technique recently to reduce the dataset size via learning a highly compact set of representative exemplars, where the model trained with these exemplars ideally should have comparable performance with respect to the one trained with the full dataset. While most of existing works upon dataset distillation focus on supervised datasets, we instead aim to distill images and their self-supervisedly trained representations into a distilled set. This procedure, named as Self-Supervised Dataset Distillation, effectively extracts rich information from real datasets, yielding the distilled sets with enhanced cross-architecture generalizability. Particularly, in order to preserve the key characteristics of original dataset more faithfully and compactly, several novel techniques are proposed: 1) we introduce an innovative parameterization upon images and representations via distinct low-dimensional bases, where the base selection for parameterization is experimentally shown to play a crucial role; 2) we tackle the instability induced by the randomness of data augmentation -- a key component in self-supervised learning but being underestimated in the prior work of self-supervised dataset distillation -- by utilizing predetermined augmentations; 3) we further leverage a lightweight network to model the connections among the representations of augmented views from the same image, leading to more compact pairs of distillation. Extensive experiments conducted on various datasets validate the superiority of our approach in terms of distillation efficiency, cross-architecture generalization, and transfer learning performance.
Authors: Saeed Ghorbani
Abstract: We introduce Aether Weaver, a novel, integrated framework for multimodal narrative co-generation that overcomes limitations of sequential text-to-visual pipelines. Our system concurrently synthesizes textual narratives, dynamic scene graph representations, visual scenes, and affective soundscapes, driven by a tightly integrated, co-generation mechanism. At its core, the Narrator, a large language model, generates narrative text and multimodal prompts, while the Director acts as a dynamic scene graph manager, and analyzes the text to build and maintain a structured representation of the story's world, ensuring spatio-temporal and relational consistency for visual rendering and subsequent narrative generation. Additionally, a Narrative Arc Controller guides the high-level story structure, influencing multimodal affective consistency, further complemented by an Affective Tone Mapper that ensures congruent emotional expression across all modalities. Through qualitative evaluations on a diverse set of narrative prompts encompassing various genres, we demonstrate that Aether Weaver significantly enhances narrative depth, visual fidelity, and emotional resonance compared to cascaded baseline approaches. This integrated framework provides a robust platform for rapid creative prototyping and immersive storytelling experiences.
Authors: Jihao Gu, Kun Li, Fei Wang, Yanyan Wei, Zhiliang Wu, Hehe Fan, Meng Wang
Abstract: Micro-Actions (MAs) are an important form of non-verbal communication in social interactions, with potential applications in human emotional analysis. However, existing methods in Micro-Action Recognition often overlook the inherent subtle changes in MAs, which limits the accuracy of distinguishing MAs with subtle changes. To address this issue, we present a novel Motion-guided Modulation Network (MMN) that implicitly captures and modulates subtle motion cues to enhance spatial-temporal representation learning. Specifically, we introduce a Motion-guided Skeletal Modulation module (MSM) to inject motion cues at the skeletal level, acting as a control signal to guide spatial representation modeling. In parallel, we design a Motion-guided Temporal Modulation module (MTM) to incorporate motion information at the frame level, facilitating the modeling of holistic motion patterns in micro-actions. Finally, we propose a motion consistency learning strategy to aggregate the motion cues from multi-scale features for micro-action classification. Experimental results on the Micro-Action 52 and iMiGUE datasets demonstrate that MMN achieves state-of-the-art performance in skeleton-based micro-action recognition, underscoring the importance of explicitly modeling subtle motion cues. The code will be available at https://github.com/momiji-bit/MMN.
Authors: Yuqi Pang, Bowen Yang, Yun Cao, Rong Fan, Xiaoyu Li, Chen He
Abstract: Vision large language models (VLLMs) are focusing primarily on handling complex and fine-grained visual information by incorporating advanced vision encoders and scaling up visual models. However, these approaches face high training and inference costs, as well as challenges in extracting visual details, effectively bridging across modalities. In this work, we propose a novel visual framework, MoCHA, to address these issues. Our framework integrates four vision backbones (i.e., CLIP, SigLIP, DINOv2 and ConvNeXt) to extract complementary visual features and is equipped with a sparse Mixture of Experts Connectors (MoECs) module to dynamically select experts tailored to different visual dimensions. To mitigate redundant or insufficient use of the visual information encoded by the MoECs module, we further design a Hierarchical Group Attention (HGA) with intra- and inter-group operations and an adaptive gating strategy for encoded visual features. We train MoCHA on two mainstream LLMs (e.g., Phi2-2.7B and Vicuna-7B) and evaluate their performance across various benchmarks. Notably, MoCHA outperforms state-of-the-art open-weight models on various tasks. For example, compared to CuMo (Mistral-7B), our MoCHA (Phi2-2.7B) presents outstanding abilities to mitigate hallucination by showing improvements of 3.25% in POPE and to follow visual instructions by raising 153 points on MME. Finally, ablation studies further confirm the effectiveness and robustness of the proposed MoECs and HGA in improving the overall performance of MoCHA.
Authors: Santosh Patapati, Trisanth Srinivasan, Murari Ambati
Abstract: Autonomous cars need geometric accuracy and semantic understanding to navigate complex environments, yet most stacks handle them separately. We present XYZ-Drive, a single vision-language model that reads a front-camera frame, a 25m $\times$ 25m overhead map, and the next waypoint, then outputs steering and speed. A lightweight goal-centered cross-attention layer lets waypoint tokens highlight relevant image and map patches, supporting both action and textual explanations, before the fused tokens enter a partially fine-tuned LLaMA-3.2 11B model. On the MD-NEX Outdoor-Driving benchmark XYZ-Drive attains 95% success and 0.80 Success weighted by Path Length (SPL), surpassing PhysNav-DG by 15%. and halving collisions, all while significantly improving efficiency by using only a single branch. Sixteen ablations explain the gains. Removing any modality (vision, waypoint, map) drops success by up to 11%, confirming their complementary roles and rich connections. Replacing goal-centered attention with simple concatenation cuts 3% in performance, showing query-based fusion injects map knowledge more effectively. Keeping the transformer frozen loses 5%, showing the importance of fine-tuning when applying VLMs for specific tasks such as autonomous driving. Coarsening map resolution from 10 cm to 40 cm blurs lane edges and raises crash rate. Overall, these results demonstrate that early, token-level fusion of intent and map layout enables accurate, transparent, real-time driving.
Authors: Henghui Ding, Song Tang, Shuting He, Chang Liu, Zuxuan Wu, Yu-Gang Jiang
Abstract: Multimodal referring segmentation aims to segment target objects in visual scenes, such as images, videos, and 3D scenes, based on referring expressions in text or audio format. This task plays a crucial role in practical applications requiring accurate object perception based on user instructions. Over the past decade, it has gained significant attention in the multimodal community, driven by advances in convolutional neural networks, transformers, and large language models, all of which have substantially improved multimodal perception capabilities. This paper provides a comprehensive survey of multimodal referring segmentation. We begin by introducing this field's background, including problem definitions and commonly used datasets. Next, we summarize a unified meta architecture for referring segmentation and review representative methods across three primary visual scenes, including images, videos, and 3D scenes. We further discuss Generalized Referring Expression (GREx) methods to address the challenges of real-world complexity, along with related tasks and practical applications. Extensive performance comparisons on standard benchmarks are also provided. We continually track related works at https://github.com/henghuiding/Awesome-Multimodal-Referring-Segmentation.
URLs: https://github.com/henghuiding/Awesome-Multimodal-Referring-Segmentation.
Authors: Chende Zheng, Ruiqi suo, Chenhao Lin, Zhengyu Zhao, Le Yang, Shuai Liu, Minghui Yang, Cong Wang, Chao Shen
Abstract: The evolution of video generation techniques, such as Sora, has made it increasingly easy to produce high-fidelity AI-generated videos, raising public concern over the dissemination of synthetic content. However, existing detection methodologies remain limited by their insufficient exploration of temporal artifacts in synthetic videos. To bridge this gap, we establish a theoretical framework through second-order dynamical analysis under Newtonian mechanics, subsequently extending the Second-order Central Difference features tailored for temporal artifact detection. Building on this theoretical foundation, we reveal a fundamental divergence in second-order feature distributions between real and AI-generated videos. Concretely, we propose Detection by Difference of Differences (D3), a novel training-free detection method that leverages the above second-order temporal discrepancies. We validate the superiority of our D3 on 4 open-source datasets (Gen-Video, VideoPhy, EvalCrafter, VidProM), 40 subsets in total. For example, on GenVideo, D3 outperforms the previous best method by 10.39% (absolute) mean Average Precision. Additional experiments on time cost and post-processing operations demonstrate D3's exceptional computational efficiency and strong robust performance. Our code is available at https://github.com/Zig-HS/D3.
Authors: Saba Ahmadi, Rabiul Awal, Ankur Sikarwar, Amirhossein Kazemnejad, Ge Ya Luo, Juan A. Rodriguez, Sai Rajeswar, Siva Reddy, Christopher Pal, Benno Krojer, Aishwarya Agrawal
Abstract: We explore three strategies to enhance performance on a wide range of image editing tasks: supervised fine-tuning (SFT), reinforcement learning (RL), and Chain-of-Thought (CoT) reasoning. In order to study all these components in one consistent framework, we adopt an autoregressive multimodal model that processes textual and visual tokens in a unified manner. We find RL combined with a large multi-modal LLM verifier to be the most effective of these strategies. As a result, we release EARL: Editing with Autoregression and RL, a strong RL-based image editing model that performs competitively on a diverse range of edits compared to strong baselines, despite using much less training data. Thus, EARL pushes the frontier of autoregressive multimodal models on image editing. We release our code, training data, and trained models at https://github.com/mair-lab/EARL.
Authors: Konstantinos Moutselos, Ilias Maglogiannis
Abstract: This study proposes a zero-shot image segmentation framework for detecting erythema (redness of the skin) using edit-friendly inversion in diffusion models. The method synthesizes reference images of the same patient that are free from erythema via generative editing and then accurately aligns these references with the original images. Color-space analysis is performed with minimal user intervention to identify erythematous regions. This approach significantly reduces the reliance on labeled dermatological datasets while providing a scalable and flexible diagnostic support tool by avoiding the need for any annotated training masks. In our initial qualitative experiments, the pipeline successfully isolated facial erythema in diverse cases, demonstrating performance improvements over baseline threshold-based techniques. These results highlight the potential of combining generative diffusion models and statistical color segmentation for computer-aided dermatology, enabling efficient erythema detection without prior training data.
Authors: Minjeong Park, Hongbeen Park, Sangwon Lee, Yoonha Jang, Jinkyu Kim
Abstract: Pedestrian Attribute Recognition (PAR) plays a crucial role in various vision tasks such as person retrieval and identification. Most existing attribute-based retrieval methods operate under the closed-set assumption that all attribute classes are consistently available during both training and inference. However, this assumption limits their applicability in real-world scenarios where novel attributes may emerge. Moreover, predefined attributes in benchmark datasets are often generic and shared across individuals, making them less discriminative for retrieving the target person. To address these challenges, we propose the Open-Attribute Recognition for Person Retrieval (OAPR) task, which aims to retrieve individuals based on attribute cues, regardless of whether those attributes were seen during training. To support this task, we introduce a novel framework designed to learn generalizable body part representations that cover a broad range of attribute categories. Furthermore, we reconstruct four widely used datasets for open-attribute recognition. Comprehensive experiments on these datasets demonstrate the necessity of the OAPR task and the effectiveness of our framework. The source code and pre-trained models will be publicly available upon publication.
Authors: Shitian Yang, Deyu Li, Xiaoke Jiang, Lei Zhang
Abstract: RGB-based 3D tasks, e.g., 3D detection, depth estimation, 3D keypoint estimation, still suffer from scarce, expensive annotations and a thin augmentation toolbox, since most image transforms, including resize and rotation, disrupt geometric consistency. In this paper, we introduce 3DRot, a plug-and-play augmentation that rotates and mirrors images about the camera's optical center while synchronously updating RGB images, camera intrinsics, object poses, and 3D annotations to preserve projective geometry-achieving geometry-consistent rotations and reflections without relying on any scene depth. We validate 3DRot with a classical 3D task, monocular 3D detection. On SUN RGB-D dataset, 3DRot raises $IoU_{3D}$ from 43.21 to 44.51, cuts rotation error (ROT) from 22.91$^\circ$ to 20.93$^\circ$, and boosts $mAP_{0.5}$ from 35.70 to 38.11. As a comparison, Cube R-CNN adds 3 other datasets together with SUN RGB-D for monocular 3D estimation, with a similar mechanism and test dataset, increases $IoU_{3D}$ from 36.2 to 37.8, boosts $mAP_{0.5}$ from 34.7 to 35.4. Because it operates purely through camera-space transforms, 3DRot is readily transferable to other 3D tasks.
Authors: Shuo Feng, Zihan Wang, Yuchen Li, Rui Kong, Hengyi Cai, Shuaiqiang Wang, Gim Hee Lee, Piji Li, Shuqiang Jiang
Abstract: While natural language is commonly used to guide embodied agents, the inherent ambiguity and verbosity of language often hinder the effectiveness of language-guided navigation in complex environments. To this end, we propose Visual Prompt Navigation (VPN), a novel paradigm that guides agents to navigate using only user-provided visual prompts within 2D top-view maps. This visual prompt primarily focuses on marking the visual navigation trajectory on a top-down view of a scene, offering intuitive and spatially grounded guidance without relying on language instructions. It is more friendly for non-expert users and reduces interpretive ambiguity. We build VPN tasks in both discrete and continuous navigation settings, constructing two new datasets, R2R-VP and R2R-CE-VP, by extending existing R2R and R2R-CE episodes with corresponding visual prompts. Furthermore, we introduce VPNet, a dedicated baseline network to handle the VPN tasks, with two data augmentation strategies: view-level augmentation (altering initial headings and prompt orientations) and trajectory-level augmentation (incorporating diverse trajectories from large-scale 3D scenes), to enhance navigation performance. Extensive experiments evaluate how visual prompt forms, top-view map formats, and data augmentation strategies affect the performance of visual prompt navigation. The code is available at https://github.com/farlit/VPN.
Authors: Florin-Alexandru Vasluianu, Tim Seizinger, Zongwei Wu, Radu Timofte
Abstract: Illumination in practical scenarios is inherently complex, involving colored light sources, occlusions, and diverse material interactions that produce intricate reflectance and shading effects. However, existing methods often oversimplify this challenge by assuming a single light source or uniform, white-balanced lighting, leaving many of these complexities unaddressed. In this paper, we introduce CL3AN, the first large-scale, high-resolution dataset of its kind designed to facilitate the restoration of images captured under multiple Colored Light sources to their Ambient-Normalized counterparts. Through benchmarking, we find that leading approaches often produce artifacts, such as illumination inconsistencies, texture leakage, and color distortion, primarily due to their limited ability to precisely disentangle illumination from reflectance. Motivated by this insight, we achieve such a desired decomposition through a novel learning framework that leverages explicit chromaticity-luminance components guidance, drawing inspiration from the principles of the Retinex model. Extensive evaluations on existing benchmarks and our dataset demonstrate the effectiveness of our approach, showcasing enhanced robustness under non-homogeneous color lighting and material-specific reflectance variations, all while maintaining a highly competitive computational cost. The benchmark, codes, and models are available at www.github.com/fvasluianu97/RLN2.
Authors: Yunuo Chen, Zezheng Lyu, Bing He, Hongwei Hu, Qi Wang, Yuan Tian, Li Song, Wenjun Zhang, Guo Lu
Abstract: Recent Learned image compression (LIC) leverages Mamba-style state-space models (SSMs) for global receptive fields with linear complexity. However, vanilla Mamba is content-agnostic, relying on fixed and predefined selective scans, which restricts its ability to dynamically and fully exploit content dependencies. We introduce Content-Adaptive Mamba (CAM), a dynamic SSM that addresses two critical limitations. First, it employs content-aware token reorganization, clustering and reordering tokens based on content similarity to prioritize proximity in feature space over Euclidean space. Second, it integrates global priors into SSM via a prompt dictionary, effectively mitigating the strict causality and long-range decay in the token interactions of Mamba. These innovations enable CAM to better capture global dependencies while preserving computational efficiency. Leveraging CAM, our Content-Adaptive Mamba-based LIC model (CMIC) achieves state-of-the-art rate-distortion performance, surpassing VTM-21.0 by -15.91\%, -21.34\%, and -17.58\% BD-rate on Kodak, Tecnick, and CLIC benchmarks, respectively.
Authors: Xiaoliu Guan, Lielin Jiang, Hanqi Chen, Xu Zhang, Jiaxing Yan, Guanzhong Wang, Yi Liu, Zetao Zhang, Yu Wu
Abstract: Diffusion Transformers (DiTs) have demonstrated remarkable performance in visual generation tasks. However, their low inference speed limits their deployment in low-resource applications. Recent training-free approaches exploit the redundancy of features across timesteps by caching and reusing past representations to accelerate inference. Building on this idea, TaylorSeer instead uses cached features to predict future ones via Taylor expansion. However, its module-level prediction across all transformer blocks (e.g., attention or feedforward modules) requires storing fine-grained intermediate features, leading to notable memory and computation overhead. Moreover, it adopts a fixed caching schedule without considering the varying accuracy of predictions across timesteps, which can lead to degraded outputs when prediction fails. To address these limitations, we propose a novel approach to better leverage Taylor-based acceleration. First, we shift the Taylor prediction target from the module level to the last block level, significantly reducing the number of cached features. Furthermore, observing strong sequential dependencies among Transformer blocks, we propose to use the error between the Taylor-estimated and actual outputs of the first block as an indicator of prediction reliability. If the error is small, we trust the Taylor prediction for the last block; otherwise, we fall back to full computation, thereby enabling a dynamic caching mechanism. Empirical results show that our method achieves a better balance between speed and quality, achieving a 3.17x acceleration on FLUX, 2.36x on DiT, and 4.14x on Wan Video with negligible quality drop. The Project Page is \href{https://cg-taylor-acce.github.io/CG-Taylor/}{here.}
Authors: Yin Lin, Riccardo Barbieri, Domenico Aquino, Giuseppe Lauria, Marina Grisoli, Elena De Momi, Alberto Redaelli, Simona Ferrante
Abstract: Glioblastoma is one of the most aggressive and common brain tumors, with a median survival of 10-15 months. Predicting Overall Survival (OS) is critical for personalizing treatment strategies and aligning clinical decisions with patient outcomes. In this study, we propose a novel Artificial Intelligence (AI) approach for OS prediction using Magnetic Resonance Imaging (MRI) images, exploiting Vision Transformers (ViTs) to extract hidden features directly from MRI images, eliminating the need of tumor segmentation. Unlike traditional approaches, our method simplifies the workflow and reduces computational resource requirements. The proposed model was evaluated on the BRATS dataset, reaching an accuracy of 62.5% on the test set, comparable to the top-performing methods. Additionally, it demonstrated balanced performance across precision, recall, and F1 score, overcoming the best model in these metrics. The dataset size limits the generalization of the ViT which typically requires larger datasets compared to convolutional neural networks. This limitation in generalization is observed across all the cited studies. This work highlights the applicability of ViTs for downsampled medical imaging tasks and establishes a foundation for OS prediction models that are computationally efficient and do not rely on segmentation.
Authors: Jianchao Wang, Peng Zhou, Cen Li, Rong Quan, Jie Qin
Abstract: 3D Gaussian Splatting (3DGS) is a powerful and computationally efficient representation for 3D reconstruction. Despite its strengths, 3DGS often produces floating artifacts, which are erroneous structures detached from the actual geometry and significantly degrade visual fidelity. The underlying mechanisms causing these artifacts, particularly in low-quality initialization scenarios, have not been fully explored. In this paper, we investigate the origins of floating artifacts from a frequency-domain perspective and identify under-optimized Gaussians as the primary source. Based on our analysis, we propose \textit{Eliminating-Floating-Artifacts} Gaussian Splatting (EFA-GS), which selectively expands under-optimized Gaussians to prioritize accurate low-frequency learning. Additionally, we introduce complementary depth-based and scale-based strategies to dynamically refine Gaussian expansion, effectively mitigating detail erosion. Extensive experiments on both synthetic and real-world datasets demonstrate that EFA-GS substantially reduces floating artifacts while preserving high-frequency details, achieving an improvement of 1.68 dB in PSNR over baseline method on our RWLQ dataset. Furthermore, we validate the effectiveness of our approach in downstream 3D editing tasks. We provide our implementation in https://jcwang-gh.github.io/EFA-GS.
Authors: Jiaze E, Srutarshi Banerjee, Tekin Bicer, Guannan Wang, Yanfu Zhang, Bin Ren
Abstract: Computed tomography (CT) is widely used in scientific and medical imaging, but acquiring full-view sinograms requires high radiation dose and long scan times. Sparse-view CT alleviates this burden but yields incomplete sinograms with structured signal loss, hampering accurate reconstruction. Unlike RGB images, sinograms encode overlapping features along projection paths and exhibit directional spectral patterns. Standard inpainting models overlook these properties, treating missing data as local holes and neglecting angular dependencies and physical consistency. We propose~\modelname, a diffusion-based framework tailored for sinograms, which restores global structure through bidirectional frequency reasoning and angular-aware masking, while enforcing physical plausibility via physics-guided constraints and frequency-adaptive noise control. Experiments on synthetic and real-world datasets show that~\modelname~consistently outperforms baselines, achieving SSIM over 0.93 and PSNR above 31 dB across diverse sparse-view scenarios.
Authors: Songning Lai, Mingqian Liao, Zhangyi Hu, Jiayu Yang, Wenshuo Chen, Hongru Xiao, Jianheng Tang, Haicheng Liao, Yutao Yue
Abstract: Concept Bottleneck Models (CBMs) enhance the interpretability of AI systems, particularly by bridging visual input with human-understandable concepts, effectively acting as a form of multimodal interpretability model. However, existing CBMs typically assume static datasets, which fundamentally limits their adaptability to real-world, continuously evolving multimodal data streams. To address this, we define a novel continual learning task for CBMs: simultaneously handling concept-incremental and class-incremental learning. This task requires models to continuously acquire new concepts (often representing cross-modal attributes) and classes while robustly preserving previously learned knowledge. To tackle this challenging problem, we propose CONceptual Continual Incremental Learning (CONCIL), a novel framework that fundamentally re-imagines concept and decision layer updates as linear regression problems. This reformulation eliminates the need for gradient-based optimization, thereby effectively preventing catastrophic forgetting. Crucially, CONCIL relies solely on recursive matrix operations, rendering it highly computationally efficient and well-suited for real-time and large-scale multimodal data applications. Experimental results compellingly demonstrate that CONCIL achieves "absolute knowledge memory" and significantly surpasses the performance of traditional CBM methods in both concept- and class-incremental settings, thus establishing a new paradigm for continual learning in CBMs, particularly valuable for dynamic multimodal understanding.
Authors: Luca Colombo, Fabrizio Pittorino, Manuel Roveri
Abstract: Binary Neural Networks (BNNs) significantly reduce computational complexity and memory usage in machine and deep learning by representing weights and activations with just one bit. However, most existing training algorithms for BNNs rely on quantization-aware floating-point Stochastic Gradient Descent (SGD), limiting the full exploitation of binary operations to the inference phase only. In this work, we propose, for the first time, a fully binary and gradient-free training algorithm for multi-layer BNNs, eliminating the need for back-propagated floating-point gradients. Specifically, the proposed algorithm relies on local binary error signals and binary weight updates, employing integer-valued hidden weights that serve as a synaptic metaplasticity mechanism, thereby enhancing its neurobiological plausibility. Our proposed solution enables the training of binary multi-layer perceptrons by using exclusively XNOR, Popcount, and increment/decrement operations. Experimental results on multi-class classification benchmarks show test accuracy improvements of up to +35.47% over the only existing fully binary single-layer state-of-the-art solution. Compared to full-precision SGD, our solution improves test accuracy by up to +35.30% under the same total memory demand, while also reducing computational cost by two to three orders of magnitude in terms of the total number of Boolean gates. The proposed algorithm is made available to the scientific community as a public repository.
Authors: Ali Tourani, Deniz Isinsu Avsar, Hriday Bavle, Jose Luis Sanchez-Lopez, Jan Lagerwall, Holger Voos
Abstract: Fiducial markers are widely used in various robotics tasks, facilitating enhanced navigation, object recognition, and scene understanding. Despite their advantages for robots and Augmented Reality (AR) applications, they often disrupt the visual aesthetics of environments because they are visible to humans, making them unsuitable for non-intrusive use cases. To address this gap, this paper presents "iMarkers"-innovative, unobtrusive fiducial markers detectable exclusively by robots equipped with specialized sensors. These markers offer high flexibility in production, allowing customization of their visibility range and encoding algorithms to suit various demands. The paper also introduces the hardware designs and software algorithms developed for detecting iMarkers, highlighting their adaptability and robustness in the detection and recognition stages. Various evaluations have demonstrated the effectiveness of iMarkers compared to conventional (printed) and blended fiducial markers and confirmed their applicability in diverse robotics scenarios.
Authors: Riccardo Ali, Francesco Caso, Christopher Irwin, Pietro Li\`o
Abstract: Transformer models map input token sequences to output token distributions, layer by layer. While most interpretability work focuses on internal latent representations, we study the evolution of these token-level distributions directly in vocabulary space. However, such distributions are high-dimensional and defined on an unordered support, making common descriptors like moments or cumulants ill-suited. We address this by computing the Shannon entropy of each intermediate predicted distribution, yielding one interpretable scalar per layer. The resulting sequence, the entropy profile, serves as a compact, information-theoretic signature of the model's computation. We introduce Entropy-Lens, a model-agnostic framework that extracts entropy profiles from frozen, off-the-shelf transformers. We show that these profiles (i) reveal family-specific computation patterns invariant under depth rescaling, (ii) are predictive of prompt type and task format, and (iii) correlate with output correctness. We further show that R\'enyi entropies yield similar results within a broad range of $\alpha$ values, justifying the use of Shannon entropy as a stable and principled summary. Our results hold across different transformers, without requiring gradients, fine-tuning, or access to model internals.
Authors: Ho Kei Cheng, Alexander Schwing
Abstract: Minibatch optimal transport coupling straightens paths in unconditional flow matching. This leads to computationally less demanding inference as fewer integration steps and less complex numerical solvers can be employed when numerically solving an ordinary differential equation at test time. However, in the conditional setting, minibatch optimal transport falls short. This is because the default optimal transport mapping disregards conditions, resulting in a conditionally skewed prior distribution during training. In contrast, at test time, we have no access to the skewed prior, and instead sample from the full, unbiased prior distribution. This gap between training and testing leads to a subpar performance. To bridge this gap, we propose conditional optimal transport C^2OT that adds a conditional weighting term in the cost matrix when computing the optimal transport assignment. Experiments demonstrate that this simple fix works with both discrete and continuous conditions in 8gaussians-to-moons, CIFAR-10, ImageNet-32x32, and ImageNet-256x256. Our method performs better overall compared to the existing baselines across different function evaluation budgets. Code is available at https://hkchengrex.github.io/C2OT
Authors: Chenxi Wang, Jizhan Fang, Xiang Chen, Bozhong Tian, Ziwen Xu, Huajun Chen, Ningyu Zhang
Abstract: Recent advancements in Large Multimodal Models (LMMs) have shown promise in Autonomous Driving Systems (ADS). However, their direct application to ADS is hindered by challenges such as misunderstanding of traffic knowledge, complex road conditions, and diverse states of vehicle. To address these challenges, we propose the use of Knowledge Editing, which enables targeted modifications to a model's behavior without the need for full retraining. Meanwhile, we introduce ADS-Edit, a multimodal knowledge editing dataset specifically designed for ADS, which includes various real-world scenarios, multiple data types, and comprehensive evaluation metrics. We conduct comprehensive experiments and derive several interesting conclusions. We hope that our work will contribute to the further advancement of knowledge editing applications in the field of autonomous driving. Code and data are available in https://github.com/zjunlp/EasyEdit/blob/main/examples/ADSEdit.md.
URLs: https://github.com/zjunlp/EasyEdit/blob/main/examples/ADSEdit.md.
Authors: Ching-Yi Lin, Sahil Shah
Abstract: Pre-trained vision transformers have achieved remarkable performance across various visual tasks but suffer from expensive computational and memory costs. While model quantization reduces memory usage by lowering precision, these models still incur significant computational overhead due to the dequantization before matrix operations. In this work, we analyze the computation graph and propose an integerization process based on operation reordering. Specifically, the process delays dequantization until after matrix operations. This enables integerized matrix multiplication and linear module by directly processing the quantized input. To validate our approach, we synthesize the self-attention module of ViT on a systolic array-based hardware. Experimental results show that our low-bit inference reduces per-PE power consumption for linear layer and matrix multiplication, bridging the gap between quantized models and efficient inference.
Authors: Chethan Krishnamurthy Ramanaik, Arjun Roy, Tobias Callies, Eirini Ntoutsi
Abstract: Adversarial robustness of deep autoencoders (AEs) remains relatively unexplored, even though their non-invertible nature poses distinct challenges. Existing attack algorithms during the optimization of imperceptible, norm-bounded adversarial perturbations to maximize output damage in AEs, often stop at sub-optimal attacks. We observe that the adversarial loss gradient vanishes when backpropagated through ill-conditioned layers. This issue arises from near-zero singular values in the Jacobians of these layers, which weaken the gradient signal during optimization. We introduce GRILL, a technique that locally restores gradient signals in ill-conditioned layers, enabling more effective norm-bounded attacks. Through extensive experiments on different architectures of popular AEs, under both sample-specific and universal attack setups, and across standard and adaptive attack settings, we show that our method significantly increases the effectiveness of our adversarial attacks, enabling a more rigorous evaluation of AE robustness.
Authors: Mario Leiva, Noel Ngu, Joshua Shay Kricheli, Aditya Taparia, Ransalu Senanayake, Paulo Shakarian, Nathaniel Bastian, John Corcoran, Gerardo Simari
Abstract: The deployment of pre-trained perception models in novel environments often leads to performance degradation due to distributional shifts. Although recent artificial intelligence approaches for metacognition use logical rules to characterize and filter model errors, improving precision often comes at the cost of reduced recall. This paper addresses the hypothesis that leveraging multiple pre-trained models can mitigate this recall reduction. We formulate the challenge of identifying and managing conflicting predictions from various models as a consistency-based abduction problem, building on the idea of abductive learning (ABL) but applying it to test-time instead of training. The input predictions and the learned error detection rules derived from each model are encoded in a logic program. We then seek an abductive explanation--a subset of model predictions--that maximizes prediction coverage while ensuring the rate of logical inconsistencies (derived from domain constraints) remains below a specified threshold. We propose two algorithms for this knowledge representation task: an exact method based on Integer Programming (IP) and an efficient Heuristic Search (HS). Through extensive experiments on a simulated aerial imagery dataset featuring controlled, complex distributional shifts, we demonstrate that our abduction-based framework outperforms individual models and standard ensemble baselines, achieving, for instance, average relative improvements of approximately 13.6\% in F1-score and 16.6\% in accuracy across 15 diverse test datasets when compared to the best individual model. Our results validate the use of consistency-based abduction as an effective mechanism to robustly integrate knowledge from multiple imperfect models in challenging, novel scenarios.
Authors: Sunil Kumar, Bowen Zhao, Leo Dirac, Paulina Varshavskaya
Abstract: Despite tremendous recent advances in large model reasoning ability, vision-language models (VLMs) still struggle with detailed visual reasoning, especially when compute resources are limited. To address this challenge, we draw inspiration from methods like Deepseek-r1 for VLMs and train smaller-scale models with Group Relative Policy Optimization (GRPO) to use external tools such as zoom. The greatest benefit is obtained with a combination of GRPO learning, a simple reward structure, a simplified tool-calling interface, allocating additional tokens to the result of the tool call, and a training data mix that over-represents visually difficult examples. Compared to similarly-sized baseline models, our method achieves better performance on some visual question-answering (VQA) tasks, thanks to the detailed visual information gathered from the external tool.
Authors: Yuansheng Li, Yunhao Zou, Linwei Chen, Ying Fu
Abstract: Interferometric Hyperspectral Imaging (IHI) is a critical technique for large-scale remote sensing tasks due to its advantages in flux and spectral resolution. However, IHI is susceptible to complex errors arising from imaging steps, and its quality is limited by existing signal processing-based reconstruction algorithms. Two key challenges hinder performance enhancement: 1) the lack of training datasets. 2) the difficulty in eliminating IHI-specific degradation components through learning-based methods. To address these challenges, we propose a novel IHI reconstruction pipeline. First, based on imaging physics and radiometric calibration data, we establish a simplified yet accurate IHI degradation model and a parameter estimation method. This model enables the synthesis of realistic IHI training datasets from hyperspectral images (HSIs), bridging the gap between IHI reconstruction and deep learning. Second, we design the Interferometric Hyperspectral Reconstruction Unfolding Transformer (IHRUT), which achieves effective spectral correction and detail restoration through a stripe-pattern enhancement mechanism and a spatial-spectral transformer architecture. Experimental results demonstrate the superior performance and generalization capability of our method.The code and are available at https://github.com/bit1120203554/IHRUT.
Authors: Melih Barsbey, Lucas Prieto, Stefanos Zafeiriou, Tolga Birdal
Abstract: Robustness and resource-efficiency are two highly desirable properties for modern machine learning models. However, achieving them jointly remains a challenge. In this paper, we identify high learning rates as a facilitator for simultaneously achieving robustness to spurious correlations and network compressibility. We demonstrate that large learning rates also produce desirable representation properties such as invariant feature utilization, class separation, and activation sparsity. Our findings indicate that large learning rates compare favorably to other hyperparameters and regularization methods, in consistently satisfying these properties in tandem. In addition to demonstrating the positive effect of large learning rates across diverse spurious correlation datasets, models, and optimizers, we also present strong evidence that the previously documented success of large learning rates in standard classification tasks is related to addressing hidden/rare spurious correlations in the training dataset. Our investigation of the mechanisms underlying this phenomenon reveals the importance of confident mispredictions of bias-conflicting samples under large learning rates.
Authors: Ran Tong, Songtao Wei, Jiaqi Liu, Lanruo Wang
Abstract: Hateful memes aimed at LGBTQ\,+ communities often evade detection by tweaking either the caption, the image, or both. We build the first robustness benchmark for this setting, pairing four realistic caption attacks with three canonical image corruptions and testing all combinations on the PrideMM dataset. Two state-of-the-art detectors, MemeCLIP and MemeBLIP2, serve as case studies, and we introduce a lightweight \textbf{Text Denoising Adapter (TDA)} to enhance the latter's resilience. Across the grid, MemeCLIP degrades more gently, while MemeBLIP2 is particularly sensitive to the caption edits that disrupt its language processing. However, the addition of the TDA not only remedies this weakness but makes MemeBLIP2 the most robust model overall. Ablations reveal that all systems lean heavily on text, but architectural choices and pre-training data significantly impact robustness. Our benchmark exposes where current multimodal safety models crack and demonstrates that targeted, lightweight modules like the TDA offer a powerful path towards stronger defences.
Authors: Yuze Wang, Yue Qi
Abstract: Recent advances in leveraging large-scale Internet photo collections for 3D reconstruction have enabled immersive virtual exploration of landmarks and historic sites worldwide. However, little attention has been given to the immersive understanding of architectural styles and structural knowledge, which remains largely confined to browsing static text-image pairs. Therefore, can we draw inspiration from 3D in-the-wild reconstruction techniques and use unconstrained photo collections to create an immersive approach for understanding the 3D structure of architectural components? To this end, we extend language embedded 3D Gaussian splatting (3DGS) and propose a novel framework for open-vocabulary scene understanding from unconstrained photo collections. Specifically, we first render multiple appearance images from the same viewpoint as the unconstrained image with the reconstructed radiance field, then extract multi-appearance CLIP features and two types of language feature uncertainty maps-transient and appearance uncertainty-derived from the multi-appearance features to guide the subsequent optimization process. Next, we propose a transient uncertainty-aware autoencoder, a multi-appearance language field 3DGS representation, and a post-ensemble strategy to effectively compress, learn, and fuse language features from multiple appearances. Finally, to quantitatively evaluate our method, we introduce PT-OVS, a new benchmark dataset for assessing open-vocabulary segmentation performance on unconstrained photo collections. Experimental results show that our method outperforms existing methods, delivering accurate open-vocabulary segmentation and enabling applications such as interactive roaming with open-vocabulary queries, architectural style pattern recognition, and 3D scene editing.
Authors: Liu Li, Qiang Ma, Cheng Ouyang, Johannes C. Paetzold, Daniel Rueckert, Bernhard Kainz
Abstract: Deep learning-based medical image segmentation techniques have shown promising results when evaluated based on conventional metrics such as the Dice score or Intersection-over-Union. However, these fully automatic methods often fail to meet clinically acceptable accuracy, especially when topological constraints should be observed, e.g., continuous boundaries or closed surfaces. In medical image segmentation, the correctness of a segmentation in terms of the required topological genus sometimes is even more important than the pixel-wise accuracy. Existing topology-aware approaches commonly estimate and constrain the topological structure via the concept of persistent homology (PH). However, these methods are difficult to implement for high dimensional data due to their polynomial computational complexity. To overcome this problem, we propose a novel and fast approach for topology-aware segmentation based on the Euler Characteristic ($\chi$). First, we propose a fast formulation for $\chi$ computation in both 2D and 3D. The scalar $\chi$ error between the prediction and ground-truth serves as the topological evaluation metric. Then we estimate the spatial topology correctness of any segmentation network via a so-called topological violation map, i.e., a detailed map that highlights regions with $\chi$ errors. Finally, the segmentation results from the arbitrary network are refined based on the topological violation maps by a topology-aware correction network. Our experiments are conducted on both 2D and 3D datasets and show that our method can significantly improve topological correctness while preserving pixel-wise segmentation accuracy.
Authors: Le Wang, Jun Wang, Feng Deng, Chen Zhang, Di Zhang, Kun Gai
Abstract: We present AudioGen-Omni - a unified approach based on multimodal diffusion transformers (MMDit), capable of generating high-fidelity audio, speech, and songs coherently synchronized with the input video. AudioGen-Omni introduces a novel joint training paradigm that seamlessly integrates large-scale video-text-audio corpora, enabling a model capable of generating semantically rich, acoustically diverse audio conditioned on multimodal inputs and adaptable to a wide range of audio generation tasks. AudioGen-Omni employs a unified lyrics-transcription encoder that encodes graphemes and phonemes from both sung and spoken inputs into dense frame-level representations. Dense frame-level representations are fused using an AdaLN-based joint attention mechanism enhanced with phase-aligned anisotropic positional infusion (PAAPI), wherein RoPE is selectively applied to temporally structured modalities to ensure precise and robust cross-modal alignment. By unfreezing all modalities and masking missing inputs, AudioGen-Omni mitigates the semantic constraints of text-frozen paradigms, enabling effective cross-modal conditioning. This joint training approach enhances audio quality, semantic alignment, and lip-sync accuracy, while also achieving state-of-the-art results on Text-to-Audio/Speech/Song tasks. With an inference time of 1.91 seconds for 8 seconds of audio, it offers substantial improvements in both efficiency and generality.
Authors: Shuangkang Fang, I-Chao Shen, Yufeng Wang, Yi-Hsuan Tsai, Yi Yang, Shuchang Zhou, Wenrui Ding, Takeo Igarashi, Ming-Hsuan Yang
Abstract: We present MeshLLM, a novel framework that leverages large language models (LLMs) to understand and generate text-serialized 3D meshes. Our approach addresses key limitations in existing methods, including the limited dataset scale when catering to LLMs' token length and the loss of 3D structural information during mesh serialization. We introduce a Primitive-Mesh decomposition strategy, which divides 3D meshes into structurally meaningful subunits. This enables the creation of a large-scale dataset with 1500k+ samples, almost 50 times larger than previous methods, which aligns better with the LLM scaling law principles. Furthermore, we propose inferring face connectivity from vertices and local mesh assembly training strategies, significantly enhancing the LLMs' ability to capture mesh topology and spatial structures. Experiments show that MeshLLM outperforms the state-of-the-art LLaMA-Mesh in both mesh generation quality and shape understanding, highlighting its great potential in processing text-serialized 3D meshes.
Authors: Sagar Singh Gwal, Rajan, Suyash Devgan, Shraddhanjali Satapathy, Abhishek Goyal, Nuruddin Mohammad Iqbal, Vivaan Jain, Prabhat Singh Mallik, Deepali Jain, Ishaan Gupta
Abstract: Lung adenocarcinoma (LUAD) is a subtype of non-small cell lung cancer (NSCLC). LUAD with mutation in the EGFR gene accounts for approximately 46% of LUAD cases. Patients carrying EGFR mutations can be treated with specific tyrosine kinase inhibitors (TKIs). Hence, predicting EGFR mutation status can help in clinical decision making. H&E-stained whole slide imaging (WSI) is a routinely performed screening procedure for cancer staging and subtyping, especially affecting the Southeast Asian populations with significantly higher incidence of the mutation when compared to Caucasians (39-64% vs 7-22%). Recent progress in AI models has shown promising results in cancer detection and classification. In this study, we propose a deep learning (DL) framework built on vision transformers (ViT) based pathology foundation model and attention-based multiple instance learning (ABMIL) architecture to predict EGFR mutation status from H&E WSI. The developed pipeline was trained using data from an Indian cohort (170 WSI) and evaluated across two independent datasets: Internal test (30 WSI from Indian cohort) set, and an external test set from TCGA (86 WSI). The model shows consistent performance across both datasets, with AUCs of 0.933 (+/-0.010), and 0.965 (+/-0.015) for the internal and external test sets respectively. This proposed framework can be efficiently trained on small datasets, achieving superior performance as compared to several prior studies irrespective of training domain. The current study demonstrates the feasibility of accurately predicting EGFR mutation status using routine pathology slides, particularly in resource-limited settings using foundation models and attention-based multiple instance learning.
Authors: Xin Ding, Yun Chen, Yongwei Wang, Kao Zhang, Sen Zhang, Peibei Cao, Xiangxue Wang
Abstract: Recent advances in conditional generative modeling have introduced Continuous conditional Generative Adversarial Network (CcGAN) and Continuous Conditional Diffusion Model (CCDM) for estimating high-dimensional data distributions conditioned on scalar, continuous regression labels (e.g., angles, ages, or temperatures). However, these approaches face fundamental limitations: CcGAN suffers from data imbalance due to fixed-size vicinity constraints, while CCDM requires computationally expensive iterative sampling. We present CcGAN-AVAR, an enhanced CcGAN framework that addresses both challenges: (1) leveraging the GAN framework's native one-step generation to overcome CCDMs' sampling bottleneck (achieving 300x-2000x faster inference), while (2) two novel components specifically target data imbalance - an adaptive vicinity mechanism that dynamically adjusts vicinity's size, and a multi-task discriminator that constructs two regularization terms (through auxiliary regression and density ratio estimation) to significantly improve generator training. Extensive experiments on four benchmark datasets (64x64 to 192x192 resolution) across eight challenging imbalanced settings demonstrate that CcGAN-AVAR achieves state-of-the-art generation quality while maintaining sampling efficiency.
Authors: Biagio Brattoli, Jack Shi, Jongchan Park, Taebum Lee, Donggeun Yoo, Sergio Pereira
Abstract: Identifying actionable driver mutations in non-small cell lung cancer (NSCLC) can impact treatment decisions and significantly improve patient outcomes. Despite guideline recommendations, broader adoption of genetic testing remains challenging due to limited availability and lengthy turnaround times. Machine Learning (ML) methods for Computational Pathology (CPath) offer a potential solution; however, research often focuses on only one or two common mutations, limiting the clinical value of these tools and the pool of patients who can benefit from them. This study evaluates various Multiple Instance Learning (MIL) techniques to detect six key actionable NSCLC driver mutations: ALK, BRAF, EGFR, ERBB2, KRAS, and MET ex14. Additionally, we introduce an Asymmetric Transformer Decoder model that employs queries and key-values of varying dimensions to maintain a low query dimensionality. This approach efficiently extracts information from patch embeddings and minimizes overfitting risks, proving highly adaptable to the MIL setting. Moreover, we present a method to directly utilize tissue type in the model, addressing a typical MIL limitation where either all regions or only some specific regions are analyzed, neglecting biological relevance. Our method outperforms top MIL models by an average of 3%, and over 4% when predicting rare mutations such as ERBB2 and BRAF, moving ML-based tests closer to being practical alternatives to standard genetic testing.
Authors: Kunyu Zhang, Lin Gu, Liangchen Liu, Yingke Chen, Binyang Wang, Jin Yan, Yingying Zhu
Abstract: Many previous studies have proposed extracting image labels from clinical notes to create large-scale medical image datasets at a low cost. However, these approaches inherently suffer from label noise due to uncertainty from the clinical experts. When radiologists and physicians analyze medical images to make diagnoses, they often include uncertainty-aware notes such as ``maybe'' or ``not excluded''. Unfortunately, current text-mining methods overlook these nuances, resulting in the creation of noisy labels. Existing methods for handling noisy labels in medical image analysis, which typically address the problem through post-processing techniques, have largely ignored the important issue of expert-driven uncertainty contributing to label noise. To better incorporate the expert-written uncertainty in clinical notes into medical image analysis and address the label noise issue, we first examine the impact of clinical expert uncertainty on label noise. We then propose a clinical expert uncertainty-aware benchmark, along with a label smoothing method, which significantly improves performance compared to current state-of-the-art approaches.