Authors: Jookyung Song, Mookyoung Kang, Nojun Kwak
Abstract: This paper presents a real-time generative drawing system that interprets and integrates both formal intent - the structural, compositional, and stylistic attributes of a sketch - and contextual intent - the semantic and thematic meaning inferred from its visual content - into a unified transformation process. Unlike conventional text-prompt-based generative systems, which primarily capture high-level contextual descriptions, our approach simultaneously analyzes ground-level intuitive geometric features such as line trajectories, proportions, and spatial arrangement, and high-level semantic cues extracted via vision-language models. These dual intent signals are jointly conditioned in a multi-stage generation pipeline that combines contour-preserving structural control with style- and content-aware image synthesis. Implemented with a touchscreen-based interface and distributed inference architecture, the system achieves low-latency, two-stage transformation while supporting multi-user collaboration on shared canvases. The resulting platform enables participants, regardless of artistic expertise, to engage in synchronous, co-authored visual creation, redefining human-AI interaction as a process of co-creation and mutual enhancement.
Authors: Chenghao Liu, Jiachen Zhang, Chengxuan Li, Zhimu Zhou, Shixin Wu, Songfang Huang, Huiling Duan
Abstract: Vision-Language-Action (VLA) models process visual inputs independently at each timestep, discarding valuable temporal information inherent in robotic manipulation tasks. This frame-by-frame processing makes models vulnerable to visual noise while ignoring the substantial coherence between consecutive frames in manipulation sequences. We propose Temporal Token Fusion (TTF), a training-free approach that intelligently integrates historical and current visual representations to enhance VLA inference quality. Our method employs dual-dimension detection combining efficient grayscale pixel difference analysis with attention-based semantic relevance assessment, enabling selective temporal token fusion through hard fusion strategies and keyframe anchoring to prevent error accumulation. Comprehensive experiments across LIBERO, SimplerEnv, and real robot tasks demonstrate consistent improvements: 4.0 percentage points average on LIBERO (72.4\% vs 68.4\% baseline), cross-environment validation on SimplerEnv (4.8\% relative improvement), and 8.7\% relative improvement on real robot tasks. Our approach proves model-agnostic, working across OpenVLA and VLA-Cache architectures. Notably, TTF reveals that selective Query matrix reuse in attention mechanisms enhances rather than compromises performance, suggesting promising directions for direct KQV matrix reuse strategies that achieve computational acceleration while improving task success rates.
Authors: Tai Inui, Steven Oh, Magdeline Kuan
Abstract: We present an unsupervised slide-quality assessment pipeline that combines seven expert-inspired visual-design metrics (whitespace, colorfulness, edge density, brightness contrast, text density, color harmony, layout balance) with CLIP-ViT embeddings, using Isolation Forest-based anomaly scoring to evaluate presentation slides. Trained on 12k professional lecture slides and evaluated on six academic talks (115 slides), our method achieved Pearson correlations up to 0.83 with human visual-quality ratings-1.79x to 3.23x stronger than scores from leading vision-language models (ChatGPT o4-mini-high, ChatGPT o3, Claude Sonnet 4, Gemini 2.5 Pro). We demonstrate convergent validity with visual ratings, discriminant validity against speaker-delivery scores, and exploratory alignment with overall impressions. Our results show that augmenting low-level design cues with multimodal embeddings closely approximates audience perceptions of slide quality, enabling scalable, objective feedback in real time.
Authors: Alexandros Gkillas, Ioulia Kapsali, Nikos Piperigkos, Aris S. Lalos
Abstract: LiDAR-based segmentation is essential for reliable perception in autonomous vehicles, yet modern segmentation networks are highly susceptible to adversarial attacks that can compromise safety. Most existing defenses are designed for networks operating directly on raw 3D point clouds and rely on large, computationally intensive generative models. However, many state-of-the-art LiDAR segmentation pipelines operate on more efficient 2D range view representations. Despite their widespread adoption, dedicated lightweight adversarial defenses for this domain remain largely unexplored. We introduce an efficient model-based purification framework tailored for adversarial defense in 2D range-view LiDAR segmentation. We propose a direct attack formulation in the range-view domain and develop an explainable purification network based on a mathematical justified optimization problem, achieving strong adversarial resilience with minimal computational overhead. Our method achieves competitive performance on open benchmarks, consistently outperforming generative and adversarial training baselines. More importantly, real-world deployment on a demo vehicle demonstrates the framework's ability to deliver accurate operation in practical autonomous driving scenarios.
Authors: Ranjan Sapkota, Manoj Karkee
Abstract: The fusion of language and vision in large vision-language models (LVLMs) has revolutionized deep learning-based object detection by enhancing adaptability, contextual reasoning, and generalization beyond traditional architectures. This in-depth review presents a structured exploration of the state-of-the-art in LVLMs, systematically organized through a three-step research review process. First, we discuss the functioning of vision language models (VLMs) for object detection, describing how these models harness natural language processing (NLP) and computer vision (CV) techniques to revolutionize object detection and localization. We then explain the architectural innovations, training paradigms, and output flexibility of recent LVLMs for object detection, highlighting how they achieve advanced contextual understanding for object detection. The review thoroughly examines the approaches used in integration of visual and textual information, demonstrating the progress made in object detection using VLMs that facilitate more sophisticated object detection and localization strategies. This review presents comprehensive visualizations demonstrating LVLMs' effectiveness in diverse scenarios including localization and segmentation, and then compares their real-time performance, adaptability, and complexity to traditional deep learning systems. Based on the review, its is expected that LVLMs will soon meet or surpass the performance of conventional methods in object detection. The review also identifies a few major limitations of the current LVLM modes, proposes solutions to address those challenges, and presents a clear roadmap for the future advancement in this field. We conclude, based on this study, that the recent advancement in LVLMs have made and will continue to make a transformative impact on object detection and robotic applications in the future.
Authors: Sauptik Dhar, Nicholas Buoncristiani, Joe Anakata, Haoyu Zhang, Michelle Munson
Abstract: The advent of large (visual) language models (LLM / LVLM) have led to a deluge of automated human-like systems in several domains including social media content generation, search and recommendation, healthcare prognosis, AI assistants for cognitive tasks etc. Although these systems have been successfully integrated in production; very little focus has been placed on sports, particularly accurate identification and natural language description of the game play. Most existing LLM/LVLMs can explain generic sports activities, but lack sufficient domain-centric sports' jargon to create natural (human-like) descriptions. This work highlights the limitations of existing SoTA LLM/LVLMs for generating production-grade sports captions from images in a desired stylized format, and proposes a two-level fine-tuned LVLM pipeline to address that. The proposed pipeline yields an improvement > 8-10% in the F1, and > 2-10% in BERT score compared to alternative approaches. In addition, it has a small runtime memory footprint and fast execution time. During Super Bowl LIX the pipeline proved its practical application for live professional sports journalism; generating highly accurate and stylized captions at the rate of 6 images per 3-5 seconds for over 1000 images during the game play.
Authors: Abu Sufian, Anirudha Ghosh, Debaditya Barman, Marco Leo, Cosimo Distante
Abstract: Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities across various downstream tasks, including biometric face recognition (FR) with description. However, demographic biases remain a critical concern in FR, as these foundation models often fail to perform equitably across diverse demographic groups, considering ethnicity/race, gender, and age. Therefore, through our work DemoBias, we conduct an empirical evaluation to investigate the extent of demographic biases in LVLMs for biometric FR with textual token generation tasks. We fine-tuned and evaluated three widely used pre-trained LVLMs: LLaVA, BLIP-2, and PaliGemma on our own generated demographic-balanced dataset. We utilize several evaluation metrics, like group-specific BERTScores and the Fairness Discrepancy Rate, to quantify and trace the performance disparities. The experimental results deliver compelling insights into the fairness and reliability of LVLMs across diverse demographic groups. Our empirical study uncovered demographic biases in LVLMs, with PaliGemma and LLaVA exhibiting higher disparities for Hispanic/Latino, Caucasian, and South Asian groups, whereas BLIP-2 demonstrated comparably consistent. Repository: https://github.com/Sufianlab/DemoBias.
Authors: Chen Chu, Cyrus Shahabi
Abstract: Spatial representation learning is essential for GeoAI applications such as urban analytics, enabling the encoding of shapes, locations, and spatial relationships (topological and distance-based) of geo-entities like points, polylines, and polygons. Existing methods either target a single geo-entity type or, like Poly2Vec, decompose entities into simpler components to enable Fourier transformation, introducing high computational cost. Moreover, since the transformed space lacks geometric alignment, these methods rely on uniform, non-adaptive sampling, which blurs fine-grained features like edges and boundaries. To address these limitations, we introduce Geo2Vec, a novel method inspired by signed distance fields (SDF) that operates directly in the original space. Geo2Vec adaptively samples points and encodes their signed distances (positive outside, negative inside), capturing geometry without decomposition. A neural network trained to approximate the SDF produces compact, geometry-aware, and unified representations for all geo-entity types. Additionally, we propose a rotation-invariant positional encoding to model high-frequency spatial variations and construct a structured and robust embedding space for downstream GeoAI models. Empirical results show that Geo2Vec consistently outperforms existing methods in representing shape and location, capturing topological and distance relationships, and achieving greater efficiency in real-world GeoAI applications. Code and Data can be found at: https://github.com/chuchen2017/GeoNeuralRepresentation.
URLs: https://github.com/chuchen2017/GeoNeuralRepresentation.
Authors: Hamza Khan
Abstract: Rice is a staple food of global importance in terms of trade, nutrition, and economic growth. Among Asian nations such as China, India, Pakistan, Thailand, Vietnam and Indonesia are leading producers of both long and short grain varieties, including basmati, jasmine, arborio, ipsala, and kainat saila. To ensure consumer satisfaction and strengthen national reputations, monitoring rice crops and grain quality is essential. Manual inspection, however, is labour intensive, time consuming and error prone, highlighting the need for automated solutions for quality control and yield improvement. This study proposes an automated approach to classify five rice grain varieties using Convolutional Neural Networks (CNN). A publicly available dataset of 75000 images was used for training and testing. Model evaluation employed accuracy, recall, precision, F1-score, ROC curves, and confusion matrices. Results demonstrated high classification accuracy with minimal misclassifications, confirming the model effectiveness in distinguishing rice varieties. In addition, an accurate diagnostic method for rice leaf diseases such as Brown Spot, Blast, Bacterial Blight, and Tungro was developed. The framework combined explainable artificial intelligence (XAI) with deep learning models including CNN, VGG16, ResNet50, and MobileNetV2. Explainability techniques such as SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) revealed how specific grain and leaf features influenced predictions, enhancing model transparency and reliability. The findings demonstrate the strong potential of deep learning in agricultural applications, paving the way for robust, interpretable systems that can support automated crop quality inspection and disease diagnosis, ultimately benefiting farmers, consumers, and the agricultural economy.
Authors: Ander Galv\'an, Marivi Higuero, Jorge Sasiain, Eduardo Jacob
Abstract: Facial recognition powered by Artificial Intelligence has achieved high accuracy in specific scenarios and applications. Nevertheless, it faces significant challenges regarding privacy and identity management, particularly when unknown individuals appear in the operational context. This paper presents the design, implementation, and evaluation of a facial recognition system within a federated learning framework tailored to open-set scenarios. The proposed approach integrates the OpenMax algorithm into federated learning, leveraging the exchange of mean activation vectors and local distance measures to reliably distinguish between known and unknown subjects. Experimental results validate the effectiveness of the proposed solution, demonstrating its potential for enhancing privacy-aware and robust facial recognition in distributed environments. -- El reconocimiento facial impulsado por Inteligencia Artificial ha demostrado una alta precisi\'on en algunos escenarios y aplicaciones. Sin embargo, presenta desaf\'ios relacionados con la privacidad y la identificaci\'on de personas, especialmente considerando que pueden aparecer sujetos desconocidos para el sistema que lo implementa. En este trabajo, se propone el dise\~no, implementaci\'on y evaluaci\'on de un sistema de reconocimiento facial en un escenario de aprendizaje federado, orientado a conjuntos abiertos. Concretamente, se dise\~na una soluci\'on basada en el algoritmo OpenMax para escenarios de aprendizaje federado. La propuesta emplea el intercambio de los vectores de activaci\'on promedio y distancias locales para identificar de manera eficaz tanto personas conocidas como desconocidas. Los experimentos realizados demuestran la implementaci\'on efectiva de la soluci\'on propuesta.
Authors: Mahdis Tourian (Centre for Environmental Intelligence, University of Exeter, Exeter, UK, Department of Computer Science, Faculty of Environment, Science and Economy, University of Exeter, Exeter, UK), Sareh Rowlands (Centre for Environmental Intelligence, University of Exeter, Exeter, UK, Department of Computer Science, Faculty of Environment, Science and Economy, University of Exeter, Exeter, UK), Remy Vandaele (Centre for Environmental Intelligence, University of Exeter, Exeter, UK, Department of Computer Science, Faculty of Environment, Science and Economy, University of Exeter, Exeter, UK), Max Fancourt (Natural England, York, UK), Rebecca Mein (Natural England, York, UK), Hywel T. P. Williams (Centre for Environmental Intelligence, University of Exeter, Exeter, UK, Department of Computer Science, Faculty of Environment, Science and Economy, University of Exeter, Exeter, UK)
Abstract: Accurate classification of terrestrial habitats is critical for biodiversity conservation, ecological monitoring, and land-use planning. Several habitat classification schemes are in use, typically based on analysis of satellite imagery with validation by field ecologists. Here we present a methodology for classification of habitats based solely on ground-level imagery (photographs), offering improved validation and the ability to classify habitats at scale (for example using citizen-science imagery). In collaboration with Natural England, a public sector organisation responsible for nature conservation in England, this study develops a classification system that applies deep learning to ground-level habitat photographs, categorising each image into one of 18 classes defined by the 'Living England' framework. Images were pre-processed using resizing, normalisation, and augmentation; re-sampling was used to balance classes in the training data and enhance model robustness. We developed and fine-tuned a DeepLabV3-ResNet101 classifier to assign a habitat class label to each photograph. Using five-fold cross-validation, the model demonstrated strong overall performance across 18 habitat classes, with accuracy and F1-scores varying between classes. Across all folds, the model achieved a mean F1-score of 0.61, with visually distinct habitats such as Bare Soil, Silt and Peat (BSSP) and Bare Sand (BS) reaching values above 0.90, and mixed or ambiguous classes scoring lower. These findings demonstrate the potential of this approach for ecological monitoring. Ground-level imagery is readily obtained, and accurate computational methods for habitat classification based on such data have many potential applications. To support use by practitioners, we also provide a simple web application that classifies uploaded images using our model.
Authors: Ming Chen, Liyuan Cui, Wenyuan Zhang, Haoxian Zhang, Yan Zhou, Xiaohan Li, Songlin Tang, Jiwen Liu, Borui Liao, Hejia Chen, Xiaoqiang Liu, Pengfei Wan
Abstract: Recently, interactive digital human video generation has attracted widespread attention and achieved remarkable progress. However, building such a practical system that can interact with diverse input signals in real time remains challenging to existing methods, which often struggle with heavy computational cost and limited controllability. In this work, we introduce an autoregressive video generation framework that enables interactive multimodal control and low-latency extrapolation in a streaming manner. With minimal modifications to a standard large language model (LLM), our framework accepts multimodal condition encodings including audio, pose, and text, and outputs spatially and semantically coherent representations to guide the denoising process of a diffusion head. To support this, we construct a large-scale dialogue dataset of approximately 20,000 hours from multiple sources, providing rich conversational scenarios for training. We further introduce a deep compression autoencoder with up to 64$\times$ reduction ratio, which effectively alleviates the long-horizon inference burden of the autoregressive model. Extensive experiments on duplex conversation, multilingual human synthesis, and interactive world model highlight the advantages of our approach in low latency, high efficiency, and fine-grained multimodal controllability.
Authors: Jefferson David Rodriguez Chivata, Davide Ghiani, Simone Maurizio La Cava, Marco Micheletto, Giulia Orr\`u, Federico Lama, Gian Luca Marcialis
Abstract: ICAO-compliant facial images, initially designed for secure biometric passports, are increasingly becoming central to identity verification in a wide range of application contexts, including border control, digital travel credentials, and financial services. While their standardization enables global interoperability, it also facilitates practices such as morphing and deepfakes, which can be exploited for harmful purposes like identity theft and illegal sharing of identity documents. Traditional countermeasures like Presentation Attack Detection (PAD) are limited to real-time capture and offer no post-capture protection. This survey paper investigates digital watermarking and steganography as complementary solutions that embed tamper-evident signals directly into the image, enabling persistent verification without compromising ICAO compliance. We provide the first comprehensive analysis of state-of-the-art techniques to evaluate the potential and drawbacks of the underlying approaches concerning the applications involving ICAO-compliant images and their suitability under standard constraints. We highlight key trade-offs, offering guidance for secure deployment in real-world identity systems.
Authors: Haoyang Su, Jin-Yi Xiang, Shaohao Rui, Yifan Gao, Xingyu Chen, Tingxuan Yin, Xiaosong Wang, Lian-Ming Wu
Abstract: Accurate prediction of major adverse cardiac events (MACE) remains a central challenge in cardiovascular prognosis. We present PRISM (Prompt-guided Representation Integration for Survival Modeling), a self-supervised framework that integrates visual representations from non-contrast cardiac cine magnetic resonance imaging with structured electronic health records (EHRs) for survival analysis. PRISM extracts temporally synchronized imaging features through motion-aware multi-view distillation and modulates them using medically informed textual prompts to enable fine-grained risk prediction. Across four independent clinical cohorts, PRISM consistently surpasses classical survival prediction models and state-of-the-art (SOTA) deep learning baselines under internal and external validation. Further clinical findings demonstrate that the combined imaging and EHR representations derived from PRISM provide valuable insights into cardiac risk across diverse cohorts. Three distinct imaging signatures associated with elevated MACE risk are uncovered, including lateral wall dyssynchrony, inferior wall hypersensitivity, and anterior elevated focus during diastole. Prompt-guided attribution further identifies hypertension, diabetes, and smoking as dominant contributors among clinical and physiological EHR factors.
Authors: Mahdieh Behjat Khatooni, Mohsen Soryani
Abstract: Alzheimer's disease (AD) is one of the most prevalent neurodegenerative disorders worldwide. As it progresses, it leads to the deterioration of cognitive functions. Since AD is irreversible, early diagnosis is crucial for managing its progression. Mild Cognitive Impairment (MCI) represents an intermediate stage between Cognitively Normal (CN) individuals and those with AD, and is considered a transitional phase from normal cognition to Alzheimer's disease. Diagnosing MCI is particularly challenging due to the subtle differences between adjacent diagnostic categories. In this study, we propose EffNetViTLoRA, a generalized end-to-end model for AD diagnosis using the whole Alzheimer's Disease Neuroimaging Initiative (ADNI) Magnetic Resonance Imaging (MRI) dataset. Our model integrates a Convolutional Neural Network (CNN) with a Vision Transformer (ViT) to capture both local and global features from MRI images. Unlike previous studies that rely on limited subsets of data, our approach is trained on the full T1-weighted MRI dataset from ADNI, resulting in a more robust and unbiased model. This comprehensive methodology enhances the model's clinical reliability. Furthermore, fine-tuning large pretrained models often yields suboptimal results when source and target dataset domains differ. To address this, we incorporate Low-Rank Adaptation (LoRA) to effectively adapt the pretrained ViT model to our target domain. This method enables efficient knowledge transfer and reduces the risk of overfitting. Our model achieves a classification accuracy of 92.52% and an F1-score of 92.76% across three diagnostic categories: AD, MCI, and CN for full ADNI dataset.
Authors: Zachary L. Crang, Rich D. Johnston, Katie L. Mills, Johsan Billingham, Sam Robertson, Michael H. Cole, Jonathon Weakley, Adam Hewitt and, Grant M. Duthie
Abstract: This study aimed to: (1) understand whether commercially available computer-vision and artificial intelligence (AI) player tracking software can accurately measure player position, speed and distance using broadcast footage and (2) determine the impact of camera feed and resolution on accuracy. Data were obtained from one match at the 2022 Qatar Federation Internationale de Football Association (FIFA) World Cup. Tactical, programme and camera 1 feeds were used. Three commercial tracking providers that use computer-vision and AI participated. Providers analysed instantaneous position (x, y coordinates) and speed (m\,s^{-1}) of each player. Their data were compared with a high-definition multi-camera tracking system (TRACAB Gen 5). Root mean square error (RMSE) and mean bias were calculated. Position RMSE ranged from 1.68 to 16.39 m, while speed RMSE ranged from 0.34 to 2.38 m\,s^{-1}. Total match distance mean bias ranged from -1745 m (-21.8%) to 1945 m (24.3%) across providers. Computer-vision and AI player tracking software offer the ability to track players with fair precision when players are detected by the software. Providers should use a tactical feed when tracking position and speed, which will maximise player detection, improving accuracy. Both 720p and 1080p resolutions are suitable, assuming appropriate computer-vision and AI models are implemented.
Authors: Xinlong Zhao, Qixiang Pang, Shan Du
Abstract: Gas leaks pose serious threats to human health and contribute significantly to atmospheric pollution, drawing increasing public concern. However, the lack of effective detection methods hampers timely and accurate identification of gas leaks. While some vision-based techniques leverage infrared videos for leak detection, the blurry and non-rigid nature of gas clouds often limits their effectiveness. To address these challenges, we propose a novel framework called Joint Vision-Language Gas leak Segmentation (JVLGS), which integrates the complementary strengths of visual and textual modalities to enhance gas leak representation and segmentation. Recognizing that gas leaks are sporadic and many video frames may contain no leak at all, our method incorporates a post-processing step to reduce false positives caused by noise and non-target objects, an issue that affects many existing approaches. Extensive experiments conducted across diverse scenarios show that JVLGS significantly outperforms state-of-the-art gas leak segmentation methods. We evaluate our model under both supervised and few-shot learning settings, and it consistently achieves strong performance in both, whereas competing methods tend to perform well in only one setting or poorly in both. Code available at: https://github.com/GeekEagle/JVLGS
Authors: Yimu Wang, Weiming Zhuang, Chen Chen, Jiabo Huang, Jingtao Li, Lingjuan Lyu
Abstract: In the era of deep learning, the increasing number of pre-trained models available online presents a wealth of knowledge. These models, developed with diverse architectures and trained on varied datasets for different tasks, provide unique interpretations of the real world. Their collective consensus is likely universal and generalizable to unseen data. However, effectively harnessing this collective knowledge poses a fundamental challenge due to the heterogeneity of pre-trained models. Existing knowledge integration solutions typically rely on strong assumptions about training data distributions and network architectures, limiting them to learning only from specific types of models and resulting in data and/or inductive biases. In this work, we introduce a novel framework, namely UNIFORM, for knowledge transfer from a diverse set of off-the-shelf models into one student model without such constraints. Specifically, we propose a dedicated voting mechanism to capture the consensus of knowledge both at the logit level -- incorporating teacher models that are capable of predicting target classes of interest -- and at the feature level, utilizing visual representations learned on arbitrary label spaces. Extensive experiments demonstrate that UNIFORM effectively enhances unsupervised object recognition performance compared to strong knowledge transfer baselines. Notably, it exhibits remarkable scalability by benefiting from over one hundred teachers, while existing methods saturate at a much smaller scale.
Authors: Xiangxu Wang, Tianhong Zhao, Wei Tu, Bowen Zhang, Guanzhou Chen, Jinzhou Cao
Abstract: Origin-Destination (OD) flow matrices are essential for urban mobility analysis, underpinning applications in traffic forecasting, infrastructure planning, and policy design. However, existing methods suffer from two critical limitations: (1) reliance on auxiliary features (e.g., Points of Interest, socioeconomic statistics) that are costly to collect and have limited spatial coverage; and (2) sensitivity to spatial topology, where minor index reordering of urban regions (e.g., census tract relabeling) disrupts structural coherence in generated flows. To address these challenges, we propose Sat2Flow, a latent structure-aware diffusion-based framework that generates structurally coherent OD flows using solely satellite imagery as input. Our approach introduces a multi-kernel encoder to capture diverse regional interactions and employs a permutation-aware diffusion process that aligns latent representations across different regional orderings. Through a joint contrastive training objective that bridges satellite-derived features with OD patterns, combined with equivariant diffusion training that enforces structural consistency, Sat2Flow ensures topological robustness under arbitrary regional reindexing. Experimental results on real-world urban datasets demonstrate that Sat2Flow outperforms both physics-based and data-driven baselines in numerical accuracy while preserving empirical distributions and spatial structures under index permutations. Sat2Flow offers a globally scalable solution for OD flow generation in data-scarce urban environments, eliminating region-specific auxiliary data dependencies while maintaining structural invariance for robust mobility modeling.
Authors: Alzayat Saleh, Shunsuke Hatano, Mostafa Rahimi Azghadi
Abstract: The automated management of invasive weeds is critical for sustainable agriculture, yet the performance of deep learning models in real-world fields is often compromised by two factors: challenging environmental conditions and the high cost of data annotation. This study tackles both issues through a diagnostic-driven, semi-supervised framework. Using a unique dataset of approximately 975 labeled and 10,000 unlabeled images of Guinea Grass in sugarcane, we first establish strong supervised baselines for classification (ResNet) and detection (YOLO, RF-DETR), achieving F1 scores up to 0.90 and mAP50 scores exceeding 0.82. Crucially, this foundational analysis, aided by interpretability tools, uncovered a pervasive "shadow bias," where models learned to misidentify shadows as vegetation. This diagnostic insight motivated our primary contribution: a semi-supervised pipeline that leverages unlabeled data to enhance model robustness. By training models on a more diverse set of visual information through pseudo-labeling, this framework not only helps mitigate the shadow bias but also provides a tangible boost in recall, a critical metric for minimizing weed escapes in automated spraying systems. To validate our methodology, we demonstrate its effectiveness in a low-data regime on a public crop-weed benchmark. Our work provides a clear and field-tested framework for developing, diagnosing, and improving robust computer vision systems for the complex realities of precision agriculture.
Authors: Zhiting Gao, Dan Song, Diqiong Jiang, Chao Xue, An-An Liu
Abstract: Motion generation is essential for animating virtual characters and embodied agents. While recent text-driven methods have made significant strides, they often struggle with achieving precise alignment between linguistic descriptions and motion semantics, as well as with the inefficiencies of slow, multi-step inference. To address these issues, we introduce TMR++ Aligned Preference Optimization (TAPO), an innovative framework that aligns subtle motion variations with textual modifiers and incorporates iterative adjustments to reinforce semantic grounding. To further enable real-time synthesis, we propose MotionFLUX, a high-speed generation framework based on deterministic rectified flow matching. Unlike traditional diffusion models, which require hundreds of denoising steps, MotionFLUX constructs optimal transport paths between noise distributions and motion spaces, facilitating real-time synthesis. The linearized probability paths reduce the need for multi-step sampling typical of sequential methods, significantly accelerating inference time without sacrificing motion quality. Experimental results demonstrate that, together, TAPO and MotionFLUX form a unified system that outperforms state-of-the-art approaches in both semantic consistency and motion quality, while also accelerating generation speed. The code and pretrained models will be released.
Authors: Nannan Zhu, Yonghao Dong, Teng Wang, Xueqian Li, Shengjun Deng, Yijia Wang, Zheng Hong, Tiantian Geng, Guo Niu, Hanyan Huang, Xiongfei Yao, Shuaiwei Jiao
Abstract: While multimodal large language models (MLLMs) exhibit strong performance on single-video tasks (e.g., video question answering), their ability across multiple videos remains critically underexplored. However, this capability is essential for real-world applications, including multi-camera surveillance and cross-video procedural learning. To bridge this gap, we present CVBench, the first comprehensive benchmark designed to assess cross-video relational reasoning rigorously. CVBench comprises 1,000 question-answer pairs spanning three hierarchical tiers: cross-video object association (identifying shared entities), cross-video event association (linking temporal or causal event chains), and cross-video complex reasoning (integrating commonsense and domain knowledge). Built from five domain-diverse video clusters (e.g., sports, life records), the benchmark challenges models to synthesise information across dynamic visual contexts. Extensive evaluation of 10+ leading MLLMs (including GPT-4o, Gemini-2.0-flash, Qwen2.5-VL) under zero-shot or chain-of-thought prompting paradigms. Key findings reveal stark performance gaps: even top models, such as GPT-4o, achieve only 60% accuracy on causal reasoning tasks, compared to the 91% accuracy of human performance. Crucially, our analysis reveals fundamental bottlenecks inherent in current MLLM architectures, notably deficient inter-video context retention and poor disambiguation of overlapping entities. CVBench establishes a rigorous framework for diagnosing and advancing multi-video reasoning, offering architectural insights for next-generation MLLMs. The data and evaluation code are available at https://github.com/Hokhim2/CVBench.
Authors: Eduardo Davalos, Yike Zhang, Namrata Srivastava, Yashvitha Thatigotla, Jorge A. Salas, Sara McFadden, Sun-Joo Cho, Amanda Goodwin, Ashwin TS, Gautam Biswas
Abstract: With advancements in AI, new gaze estimation methods are exceeding state-of-the-art (SOTA) benchmarks, but their real-world application reveals a gap with commercial eye-tracking solutions. Factors like model size, inference time, and privacy often go unaddressed. Meanwhile, webcam-based eye-tracking methods lack sufficient accuracy, in particular due to head movement. To tackle these issues, we introduce We bEyeTrack, a framework that integrates lightweight SOTA gaze estimation models directly in the browser. It incorporates model-based head pose estimation and on-device few-shot learning with as few as nine calibration samples (k < 9). WebEyeTrack adapts to new users, achieving SOTA performance with an error margin of 2.32 cm on GazeCapture and real-time inference speeds of 2.4 milliseconds on an iPhone 14. Our open-source code is available at https://github.com/RedForestAi/WebEyeTrack.
Authors: Yu-Wei Zhang, Tongju Han, Lipeng Gao, Mingqiang Wei, Hui Liu, Changbao Li, Caiming Zhang
Abstract: This paper presents MonoRelief V2, an end-to-end model designed for directly recovering 2.5D reliefs from single images under complex material and illumination variations. In contrast to its predecessor, MonoRelief V1 [1], which was solely trained on synthetic data, MonoRelief V2 incorporates real data to achieve improved robustness, accuracy and efficiency. To overcome the challenge of acquiring large-scale real-world dataset, we generate approximately 15,000 pseudo real images using a text-to-image generative model, and derive corresponding depth pseudo-labels through fusion of depth and normal predictions. Furthermore, we construct a small-scale real-world dataset (800 samples) via multi-view reconstruction and detail refinement. MonoRelief V2 is then progressively trained on the pseudo-real and real-world datasets. Comprehensive experiments demonstrate its state-of-the-art performance both in depth and normal predictions, highlighting its strong potential for a range of downstream applications. Code is at: https://github.com/glp1001/MonoreliefV2.
Authors: Yuhang Zhao, Zixing Wang
Abstract: End-to-end object detectors offer a promising NMS-free paradigm for real-time applications, yet their high computational cost remains a significant barrier, particularly for complex scenarios like intersection traffic monitoring. To address this challenge, we propose FlowDet, a high-speed detector featuring a decoupled encoder optimization strategy applied to the DETR architecture. Specifically, FlowDet employs a novel Geometric Deformable Unit (GDU) for traffic-aware geometric modeling and a Scale-Aware Attention (SAA) module to maintain high representational power across extreme scale variations. To rigorously evaluate the model's performance in environments with severe occlusion and high object density, we collected the Intersection-Flow-5k dataset, a new challenging scene for this task. Evaluated on Intersection-Flow-5k, FlowDet establishes a new state-of-the-art. Compared to the strong RT-DETR baseline, it improves AP(test) by 1.5% and AP50(test) by 1.6%, while simultaneously reducing GFLOPs by 63.2% and increasing inference speed by 16.2%. Our work demonstrates a new path towards building highly efficient and accurate detectors for demanding, real-world perception systems. The Intersection-Flow-5k dataset is available at https://github.com/AstronZh/Intersection-Flow-5K.
Authors: Luhu Li, Bowen Lin, Mukhtiar Khan, Shujun Fu
Abstract: Anomaly detection in medical images is challenging due to limited annotations and a domain gap compared to natural images. Existing reconstruction methods often rely on frozen pre-trained encoders, which limits adaptation to domain-specific features and reduces localization accuracy. Prototype-based learning offers interpretability and clustering benefits but suffers from prototype collapse, where few prototypes dominate training, harming diversity and generalization. To address this, we propose a unified framework combining a trainable encoder with prototype-guided reconstruction and a novel Diversity-Aware Alignment Loss. The trainable encoder, enhanced by a momentum branch, enables stable domain-adaptive feature learning. A lightweight Prototype Extractor mines informative normal prototypes to guide the decoder via attention for precise reconstruction. Our loss enforces balanced prototype use through diversity constraints and per-prototype normalization, effectively preventing collapse. Experiments on multiple medical imaging benchmarks show significant improvements in representation quality and anomaly localization, outperforming prior methods. Visualizations and prototype assignment analyses further validate the effectiveness of our anti-collapse mechanism and enhanced interpretability.
Authors: Mingxi Fu, Fanglei Fu, Xitong Ling, Huaitian Yuan, Tian Guan, Yonghong He, Lianghui Zhu
Abstract: Pathological image segmentation faces numerous challenges, particularly due to ambiguous semantic boundaries and the high cost of pixel-level annotations. Although recent semi-supervised methods based on consistency regularization (e.g., UniMatch) have made notable progress, they mainly rely on perturbation-based consistency within the image modality, making it difficult to capture high-level semantic priors, especially in structurally complex pathology images. To address these limitations, we propose MPAMatch - a novel segmentation framework that performs pixel-level contrastive learning under a multimodal prototype-guided supervision paradigm. The core innovation of MPAMatch lies in the dual contrastive learning scheme between image prototypes and pixel labels, and between text prototypes and pixel labels, providing supervision at both structural and semantic levels. This coarse-to-fine supervisory strategy not only enhances the discriminative capability on unlabeled samples but also introduces the text prototype supervision into segmentation for the first time, significantly improving semantic boundary modeling. In addition, we reconstruct the classic segmentation architecture (TransUNet) by replacing its ViT backbone with a pathology-pretrained foundation model (Uni), enabling more effective extraction of pathology-relevant features. Extensive experiments on GLAS, EBHI-SEG-GLAND, EBHI-SEG-CANCER, and KPI show MPAMatch's superiority over state-of-the-art methods, validating its dual advantages in structural and semantic modeling.
Authors: Zhu Xu, Zhaowen Wang, Yuxin Peng, Yang Liu
Abstract: Compositional Customized Image Generation aims to customize multiple target concepts within generation content, which has gained attention for its wild application. Existing approaches mainly concentrate on the target entity's appearance preservation, while neglecting the fine-grained interaction control among target entities. To enable the model of such interaction control capability, we focus on human object interaction scenario and propose the task of Customized Human Object Interaction Image Generation(CHOI), which simultaneously requires identity preservation for target human object and the interaction semantic control between them. Two primary challenges exist for CHOI:(1)simultaneous identity preservation and interaction control demands require the model to decompose the human object into self-contained identity features and pose-oriented interaction features, while the current HOI image datasets fail to provide ideal samples for such feature-decomposed learning.(2)inappropriate spatial configuration between human and object may lead to the lack of desired interaction semantics. To tackle it, we first process a large-scale dataset, where each sample encompasses the same pair of human object involving different interactive poses. Then we design a two-stage model Interact-Custom, which firstly explicitly models the spatial configuration by generating a foreground mask depicting the interaction behavior, then under the guidance of this mask, we generate the target human object interacting while preserving their identities features. Furthermore, if the background image and the union location of where the target human object should appear are provided by users, Interact-Custom also provides the optional functionality to specify them, offering high content controllability. Extensive experiments on our tailored metrics for CHOI task demonstrate the effectiveness of our approach.
Authors: Haomiao Zhang, Miao Cao, Xuan Yu, Hui Luo, Yanling Piao, Mengjie Qin, Zhangyuan Li, Ping Wang, Xin Yuan
Abstract: Computer-generated holography (CGH) is a promising technology for next-generation displays. However, generating high-speed, high-quality holographic video requires both high frame rate display and efficient computation, but is constrained by two key limitations: ($i$) Learning-based models often produce over-smoothed phases with narrow angular spectra, causing severe color crosstalk in high frame rate full-color displays such as depth-division multiplexing and thus resulting in a trade-off between frame rate and color fidelity. ($ii$) Existing frame-by-frame optimization methods typically optimize frames independently, neglecting spatial-temporal correlations between consecutive frames and leading to computationally inefficient solutions. To overcome these challenges, in this paper, we propose a novel high-speed full-color video CGH generation scheme. First, we introduce Spectrum-Guided Depth Division Multiplexing (SGDDM), which optimizes phase distributions via frequency modulation, enabling high-fidelity full-color display at high frame rates. Second, we present HoloMamba, a lightweight asymmetric Mamba-Unet architecture that explicitly models spatial-temporal correlations across video sequences to enhance reconstruction quality and computational efficiency. Extensive simulated and real-world experiments demonstrate that SGDDM achieves high-fidelity full-color display without compromise in frame rate, while HoloMamba generates FHD (1080p) full-color holographic video at over 260 FPS, more than 2.6$\times$ faster than the prior state-of-the-art Divide-Conquer-and-Merge Strategy.
Authors: Dat Nguyen Cong, Hieu Tran Bao, Hoang Thanh-Tung
Abstract: Diffusion models have gained prominence as state-of-the-art techniques for synthesizing images and videos, particularly due to their ability to scale effectively with large datasets. Recent studies have uncovered that these extensive datasets often contain mistakes from manual labeling processes. However, the extent to which such errors compromise the generative capabilities and controllability of diffusion models is not well studied. This paper introduces Score-based Discriminator Correction (SBDC), a guidance technique for aligning noisy pre-trained conditional diffusion models. The guidance is built on discriminator training using adversarial loss, drawing on prior noise detection techniques to assess the authenticity of each sample. We further show that limiting the usage of our guidance to the early phase of the generation process leads to better performance. Our method is computationally efficient, only marginally increases inference time, and does not require retraining diffusion models. Experiments on different noise settings demonstrate the superiority of our method over previous state-of-the-art methods.
Authors: Abhinav Kumar
Abstract: Monocular 3D object detection (Mono3D) is a fundamental computer vision task that estimates an object's class, 3D position, dimensions, and orientation from a single image. Its applications, including autonomous driving, augmented reality, and robotics, critically rely on accurate 3D environmental understanding. This thesis addresses the challenge of generalizing Mono3D models to diverse scenarios, including occlusions, datasets, object sizes, and camera parameters. To enhance occlusion robustness, we propose a mathematically differentiable NMS (GrooMeD-NMS). To improve generalization to new datasets, we explore depth equivariant (DEVIANT) backbones. We address the issue of large object detection, demonstrating that it's not solely a data imbalance or receptive field problem but also a noise sensitivity issue. To mitigate this, we introduce a segmentation-based approach in bird's-eye view with dice loss (SeaBird). Finally, we mathematically analyze the extrapolation of Mono3D models to unseen camera heights and improve Mono3D generalization in such out-of-distribution settings.
Authors: Toghrul Karimov, Hassan Imani, Allan Kazakov
Abstract: Post-training quantization (PTQ) is crucial for deploying efficient object detection models, like YOLO, on resource-constrained devices. However, the impact of reduced precision on model robustness to real-world input degradations such as noise, blur, and compression artifacts is a significant concern. This paper presents a comprehensive empirical study evaluating the robustness of YOLO models (nano to extra-large scales) across multiple precision formats: FP32, FP16 (TensorRT), Dynamic UINT8 (ONNX), and Static INT8 (TensorRT). We introduce and evaluate a degradation-aware calibration strategy for Static INT8 PTQ, where the TensorRT calibration process is exposed to a mix of clean and synthetically degraded images. Models were benchmarked on the COCO dataset under seven distinct degradation conditions (including various types and levels of noise, blur, low contrast, and JPEG compression) and a mixed-degradation scenario. Results indicate that while Static INT8 TensorRT engines offer substantial speedups (~1.5-3.3x) with a moderate accuracy drop (~3-7% mAP50-95) on clean data, the proposed degradation-aware calibration did not yield consistent, broad improvements in robustness over standard clean-data calibration across most models and degradations. A notable exception was observed for larger model scales under specific noise conditions, suggesting model capacity may influence the efficacy of this calibration approach. These findings highlight the challenges in enhancing PTQ robustness and provide insights for deploying quantized detectors in uncontrolled environments. All code and evaluation tables are available at https://github.com/AllanK24/QRID.
Authors: Qizhe Fan, Chaoyu Liu, Zhonghua Qiao, Xiaoqin Shen
Abstract: Domain Generalized Semantic Segmentation (DGSS) focuses on training a model using labeled data from a source domain, with the goal of achieving robust generalization to unseen target domains during inference. A common approach to improve generalization is to augment the source domain with synthetic data generated by diffusion models (DMs). However, the generated images often contain structural or semantic defects due to training imperfections. Training segmentation models with such flawed data can lead to performance degradation and error accumulation. To address this issue, we propose to integrate inverse evolution layers (IELs) into the generative process. IELs are designed to highlight spatial discontinuities and semantic inconsistencies using Laplacian-based priors, enabling more effective filtering of undesirable generative patterns. Based on this mechanism, we introduce IELDM, an enhanced diffusion-based data augmentation framework that can produce higher-quality images. Furthermore, we observe that the defect-suppression capability of IELs can also benefit the segmentation network by suppressing artifact propagation. Based on this insight, we embed IELs into the decoder of the DGSS model and propose IELFormer to strengthen generalization capability in cross-domain scenarios. To further strengthen the model's semantic consistency across scales, IELFormer incorporates a multi-scale frequency fusion (MFF) module, which performs frequency-domain analysis to achieve structured integration of multi-resolution features, thereby improving cross-scale coherence. Extensive experiments on benchmark datasets demonstrate that our approach achieves superior generalization performance compared to existing methods.
Authors: Jiajun Sun, Zhen Yu, Siyuan Yan, Jason J. Ong, Zongyuan Ge, Lei Zhang
Abstract: Skin images from real-world clinical practice are often limited, resulting in a shortage of training data for deep-learning models. While many studies have explored skin image synthesis, existing methods often generate low-quality images and lack control over the lesion's location and type. To address these limitations, we present LF-VAR, a model leveraging quantified lesion measurement scores and lesion type labels to guide the clinically relevant and controllable synthesis of skin images. It enables controlled skin synthesis with specific lesion characteristics based on language prompts. We train a multiscale lesion-focused Vector Quantised Variational Auto-Encoder (VQVAE) to encode images into discrete latent representations for structured tokenization. Then, a Visual AutoRegressive (VAR) Transformer trained on tokenized representations facilitates image synthesis. Lesion measurement from the lesion region and types as conditional embeddings are integrated to enhance synthesis fidelity. Our method achieves the best overall FID score (average 0.74) among seven lesion types, improving upon the previous state-of-the-art (SOTA) by 6.3%. The study highlights our controllable skin synthesis model's effectiveness in generating high-fidelity, clinically relevant synthetic skin images. Our framework code is available at https://github.com/echosun1996/LF-VAR.
Authors: Xiaolei Wei, Yi Ouyang, Haibo Ye
Abstract: Long-tailed visual recognition is challenging not only due to class imbalance but also because of varying classification difficulty across categories. Simply reweighting classes by frequency often overlooks those that are intrinsically hard to learn. To address this, we propose \textbf{DQRoute}, a modular framework that combines difficulty-aware optimization with dynamic expert collaboration. DQRoute first estimates class-wise difficulty based on prediction uncertainty and historical performance, and uses this signal to guide training with adaptive loss weighting. On the architectural side, DQRoute employs a mixture-of-experts design, where each expert specializes in a different region of the class distribution. At inference time, expert predictions are weighted by confidence scores derived from expert-specific OOD detectors, enabling input-adaptive routing without the need for a centralized router. All components are trained jointly in an end-to-end manner. Experiments on standard long-tailed benchmarks demonstrate that DQRoute significantly improves performance, particularly on rare and difficult classes, highlighting the benefit of integrating difficulty modeling with decentralized expert routing.
Authors: Yang Li, Quan Yuan, Guiyang Luo, Xiaoyuan Fu, Rui Pan, Yujia Yang, Congzhang Shao, Yuewen Liu, Jinglin Li
Abstract: Collaborative perception allows agents to enhance their perceptual capabilities by exchanging intermediate features. Existing methods typically organize these intermediate features as 2D bird's-eye-view (BEV) representations, which discard critical fine-grained 3D structural cues essential for accurate object recognition and localization. To this end, we first introduce point-level tokens as intermediate representations for collaborative perception. However, point-cloud data are inherently unordered, massive, and position-sensitive, making it challenging to produce compact and aligned point-level token sequences that preserve detailed structural information. Therefore, we present CoPLOT, a novel Collaborative perception framework that utilizes Point-Level Optimized Tokens. It incorporates a point-native processing pipeline, including token reordering, sequence modeling, and multi-agent spatial alignment. A semantic-aware token reordering module generates adaptive 1D reorderings by leveraging scene-level and token-level semantic information. A frequency-enhanced state space model captures long-range sequence dependencies across both spatial and spectral domains, improving the differentiation between foreground tokens and background clutter. Lastly, a neighbor-to-ego alignment module applies a closed-loop process, combining global agent-level correction with local token-level refinement to mitigate localization noise. Extensive experiments on both simulated and real-world datasets show that CoPLOT outperforms state-of-the-art models, with even lower communication and computation overhead. Code will be available at https://github.com/CheeryLeeyy/CoPLOT.
Authors: Bikash Kumar Badatya, Vipul Baghel, Ravi Hegde
Abstract: Fine-grained action localization in untrimmed sports videos presents a significant challenge due to rapid and subtle motion transitions over short durations. Existing supervised and weakly supervised solutions often rely on extensive annotated datasets and high-capacity models, making them computationally intensive and less adaptable to real-world scenarios. In this work, we introduce a lightweight and unsupervised skeleton-based action localization pipeline that leverages spatio-temporal graph neural representations. Our approach pre-trains an Attention-based Spatio-Temporal Graph Convolutional Network (ASTGCN) on a pose-sequence denoising task with blockwise partitions, enabling it to learn intrinsic motion dynamics without any manual labeling. At inference, we define a novel Action Dynamics Metric (ADM), computed directly from low-dimensional ASTGCN embeddings, which detects motion boundaries by identifying inflection points in its curvature profile. Our method achieves a mean Average Precision (mAP) of 82.66% and average localization latency of 29.09 ms on the DSV Diving dataset, matching state-of-the-art supervised performance while maintaining computational efficiency. Furthermore, it generalizes robustly to unseen, in-the-wild diving footage without retraining, demonstrating its practical applicability for lightweight, real-time action analysis systems in embedded or dynamic environments.
Authors: Dongjin Kim, Jaekyun Ko, Muhammad Kashif Ali, Tae Hyun Kim
Abstract: Image denoising is a fundamental challenge in computer vision, with applications in photography and medical imaging. While deep learning-based methods have shown remarkable success, their reliance on specific noise distributions limits generalization to unseen noise types and levels. Existing approaches attempt to address this with extensive training data and high computational resources but they still suffer from overfitting. To address these issues, we conduct image denoising by utilizing dynamically generated kernels via efficient operations. This approach helps prevent overfitting and improves resilience to unseen noise. Specifically, our method leverages a Feature Extraction Module for robust noise-invariant features, Global Statistics and Local Correlation Modules to capture comprehensive noise characteristics and structural correlations. The Kernel Prediction Module then employs these cues to produce pixel-wise varying kernels adapted to local structures, which are then applied iteratively for denoising. This ensures both efficiency and superior restoration quality. Despite being trained on single-level Gaussian noise, our compact model (~ 0.04 M) excels across diverse noise types and levels, demonstrating the promise of iterative dynamic filtering for practical image denoising.
Authors: Hou Xia, Zheren Fu, Fangcan Ling, Jiajun Li, Yi Tu, Zhendong Mao, Yongdong Zhang
Abstract: Large video language models (LVLMs) have made notable progress in video understanding, spurring the development of corresponding evaluation benchmarks. However, existing benchmarks generally assess overall performance across entire video sequences, overlooking nuanced behaviors such as contextual positional bias, a critical yet under-explored aspect of LVLM performance. We present Video-LevelGauge, a dedicated benchmark designed to systematically assess positional bias in LVLMs. We employ standardized probes and customized contextual setups, allowing flexible control over context length, probe position, and contextual types to simulate diverse real-world scenarios. In addition, we introduce a comprehensive analysis method that combines statistical measures with morphological pattern recognition to characterize bias. Our benchmark comprises 438 manually curated videos spanning multiple types, yielding 1,177 high-quality multiple-choice questions and 120 open-ended questions, validated for their effectiveness in exposing positional bias. Based on these, we evaluate 27 state-of-the-art LVLMs, including both commercial and open-source models. Our findings reveal significant positional biases in many leading open-source models, typically exhibiting head or neighbor-content preferences. In contrast, commercial models such as Gemini2.5-Pro show impressive, consistent performance across entire video sequences. Further analyses on context length, context variation, and model scale provide actionable insights for mitigating bias and guiding model enhancement.https://github.com/Cola-any/Video-LevelGauge
Authors: B\'alint M\'esz\'aros, Ahmet Firintepe, Sebastian Schmidt, Stephan G\"unnemann
Abstract: AI tasks in the car interior like identifying and localizing externally introduced objects is crucial for response quality of personal assistants. However, computational resources of on-board systems remain highly constrained, restricting the deployment of such solutions directly within the vehicle. To address this limitation, we propose the novel Object Detection and Localization (ODAL) framework for interior scene understanding. Our approach leverages vision foundation models through a distributed architecture, splitting computational tasks between on-board and cloud. This design overcomes the resource constraints of running foundation models directly in the car. To benchmark model performance, we introduce ODALbench, a new metric for comprehensive assessment of detection and localization.Our analysis demonstrates the framework's potential to establish new standards in this domain. We compare the state-of-the-art GPT-4o vision foundation model with the lightweight LLaVA 1.5 7B model and explore how fine-tuning enhances the lightweight models performance. Remarkably, our fine-tuned ODAL-LLaVA model achieves an ODAL$_{score}$ of 89%, representing a 71% improvement over its baseline performance and outperforming GPT-4o by nearly 20%. Furthermore, the fine-tuned model maintains high detection accuracy while significantly reducing hallucinations, achieving an ODAL$_{SNR}$ three times higher than GPT-4o.
Authors: Zongxia Li, Wenhao Yu, Chengsong Huang, Rui Liu, Zhenwen Liang, Fuxiao Liu, Jingxi Che, Dian Yu, Jordan Boyd-Graber, Haitao Mi, Dong Yu
Abstract: Vision-Language Models (VLMs) often suffer from visual hallucinations, saying things that are not actually in the image, and language shortcuts, where they skip the visual part and just rely on text priors. These issues arise because most post-training methods for VLMs rely on simple verifiable answer matching and supervise only final outputs, leaving intermediate visual reasoning without explicit guidance. As a result, VLMs receive sparse visual signals and often learn to prioritize language-based reasoning over visual perception. To mitigate this, some existing methods add visual supervision using human annotations or distilled labels from external large models. However, human annotations are labor-intensive and costly, and because external signals cannot adapt to the evolving policy, they cause distributional shifts that can lead to reward hacking. In this paper, we introduce Vision-SR1, a self-rewarding method that improves visual reasoning without relying on external visual supervisions via reinforcement learning. Vision-SR1 decomposes VLM reasoning into two stages: visual perception and language reasoning. The model is first prompted to produce self-contained visual perceptions that are sufficient to answer the question without referring back the input image. To validate this self-containment, the same VLM model is then re-prompted to perform language reasoning using only the generated perception as input to compute reward. This self-reward is combined with supervision on final outputs, providing a balanced training signal that strengthens both visual perception and language reasoning. Our experiments demonstrate that Vision-SR1 improves visual reasoning, mitigates visual hallucinations, and reduces reliance on language shortcuts across diverse vision-language tasks.
Authors: Matthias H\"offlin, J\"urgen Wassner
Abstract: Spiking Neural Networks (SNNs), inspired by biological intelligence, have long been considered inherently energy-efficient, making them attractive for resource-constrained domains such as space applications. However, recent comparative studies with conventional Artificial Neural Networks (ANNs) have begun to question this reputation, especially for digital implementations. This work investigates SNNs for multi-output regression, specifically 3-D satellite position estimation from monocular images, and compares hardware-aware and hardware-agnostic energy estimation methods. The proposed SNN, trained using the membrane potential of the Leaky Integrate-and-Fire (LIF) neuron in the final layer, achieves comparable Mean Squared Error (MSE) to a reference Convolutional Neural Network (CNN) on a photorealistic satellite dataset. Energy analysis shows that while hardware-agnostic methods predict a consistent 50-60% energy advantage for SNNs over CNNs, hardware-aware analysis reveals that significant energy savings are realized only on neuromorphic hardware and with high input sparsity. The influence of dark pixel ratio on energy consumption is quantified, emphasizing the impact of data characteristics and hardware assumptions. These findings highlight the need for transparent evaluation methods and explicit disclosure of underlying assumptions to ensure fair comparisons of neural network energy efficiency.
Authors: Weicheng Liao, Zan Chen, Jianyang Xie, Yalin Zheng, Yuhui Ma, Yitian Zhao
Abstract: Ultra-Wide-Field (UWF) retinal imaging has revolutionized retinal diagnostics by providing a comprehensive view of the retina. However, it often suffers from quality-degrading factors such as blurring and uneven illumination, which obscure fine details and mask pathological information. While numerous retinal image enhancement methods have been proposed for other fundus imageries, they often fail to address the unique requirements in UWF, particularly the need to preserve pathological details. In this paper, we propose a novel frequency-aware self-supervised learning method for UWF image enhancement. It incorporates frequency-decoupled image deblurring and Retinex-guided illumination compensation modules. An asymmetric channel integration operation is introduced in the former module, so as to combine global and local views by leveraging high- and low-frequency information, ensuring the preservation of fine and broader structural details. In addition, a color preservation unit is proposed in the latter Retinex-based module, to provide multi-scale spatial and frequency information, enabling accurate illumination estimation and correction. Experimental results demonstrate that the proposed work not only enhances visualization quality but also improves disease diagnosis performance by restoring and correcting fine local details and uneven intensity. To the best of our knowledge, this work is the first attempt for UWF image enhancement, offering a robust and clinically valuable tool for improving retinal disease management.
Authors: Gangjian Zhang, Jian Shu, Nanjie Yao, Hao Wang
Abstract: Monocular texture 3D human reconstruction aims to create a complete 3D digital avatar from just a single front-view human RGB image. However, the geometric ambiguity inherent in a single 2D image and the scarcity of 3D human training data are the main obstacles limiting progress in this field. To address these issues, current methods employ prior geometric estimation networks to derive various human geometric forms, such as the SMPL model and normal maps. However, they struggle to integrate these modalities effectively, leading to view inconsistencies, such as facial distortions. To this end, we propose a two-process 3D human reconstruction framework, SAT, which seamlessly learns various prior geometries in a unified manner and reconstructs high-quality textured 3D avatars as the final output. To further facilitate geometry learning, we introduce a Supervisor Feature Regularization module. By employing a multi-view network with the same structure to provide intermediate features as training supervision, these varied geometric priors can be better fused. To tackle data scarcity and further improve reconstruction quality, we also propose an Online Animation Augmentation module. By building a one-feed-forward animation network, we augment a massive number of samples from the original 3D human data online for model training. Extensive experiments on two benchmarks show the superiority of our approach compared to state-of-the-art methods.
Authors: V. S. Usatyuk, D. A. Sapozhnikov, S. I. Egorov
Abstract: The rapid advance of deep generative models such as GANs and diffusion networks now produces images that are virtually indistinguishable from genuine photographs, undermining media forensics and biometric security. Supervised detectors quickly lose effectiveness on unseen generators or after adversarial post-processing, while existing unsupervised methods that rely on low-level statistical cues remain fragile. We introduce a physics-inspired, model-agnostic detector that treats synthetic-image identification as a community-detection problem on a sparse weighted graph. Image features are first extracted with pretrained CNNs and reduced to 32 dimensions, each feature vector becomes a node of a Multi-Edge Type QC-LDPC graph. Pairwise similarities are transformed into edge couplings calibrated at the Nishimori temperature, producing a Random Bond Ising Model (RBIM) whose Bethe-Hessian spectrum exhibits a characteristic gap when genuine community structure (real images) is present. Synthetic images violate the Nishimori symmetry and therefore lack such gaps. We validate the approach on binary tasks cat versus dog and male versus female using real photos from Flickr-Faces-HQ and CelebA and synthetic counterparts generated by GANs and diffusion models. Without any labeled synthetic data or retraining of the feature extractor, the detector achieves over 94% accuracy. Spectral analysis shows multiple well separated gaps for real image sets and a collapsed spectrum for generated ones. Our contributions are threefold: a novel LDPC graph construction that embeds deep image features, an analytical link between Nishimori temperature RBIM and the Bethe-Hessian spectrum providing a Bayes optimal detection criterion; and a practical, unsupervised synthetic image detector robust to new generative architectures. Future work will extend the framework to video streams and multi-class anomaly detection.
Authors: Yupeng Zhang, Dezhi Zheng, Ping Lu, Han Zhang, Lei Wang, Liping xiang, Cheng Luo, Kaijun Deng, Xiaowen Fu, Linlin Shen, Jinbao Wang
Abstract: 3D Gaussian Splatting (3DGS) has emerged as a novel explicit representation for 3D scenes, offering both high-fidelity reconstruction and efficient rendering. However, 3DGS lacks 3D segmentation ability, which limits its applicability in tasks that require scene understanding. The identification and isolating of specific object components is crucial. To address this limitation, we propose Label-aware 3D Gaussian Splatting (LabelGS), a method that augments the Gaussian representation with object label.LabelGS introduces cross-view consistent semantic masks for 3D Gaussians and employs a novel Occlusion Analysis Model to avoid overfitting occlusion during optimization, Main Gaussian Labeling model to lift 2D semantic prior to 3D Gaussian and Gaussian Projection Filter to avoid Gaussian label conflict. Our approach achieves effective decoupling of Gaussian representations and refines the 3DGS optimization process through a random region sampling strategy, significantly improving efficiency. Extensive experiments demonstrate that LabelGS outperforms previous state-of-the-art methods, including Feature-3DGS, in the 3D scene segmentation task. Notably, LabelGS achieves a remarkable 22X speedup in training compared to Feature-3DGS, at a resolution of 1440X1080. Our code will be at https://github.com/garrisonz/LabelGS.
Authors: Qiang Hu, Ying Zhou, Gepeng Ji, Nick Barnes, Qiang Li, Zhiwei Wang
Abstract: Existing video polyp segmentation (VPS) paradigms usually struggle to balance between spatiotemporal modeling and domain generalization, limiting their applicability in real clinical scenarios. To embrace this challenge, we recast the VPS task as a track-by-detect paradigm that leverages the spatial contexts captured by the image polyp segmentation (IPS) model while integrating the temporal modeling capabilities of segment anything model 2 (SAM2). However, during long-term polyp tracking in colonoscopy videos, SAM2 suffers from error accumulation, resulting in a snowball effect that compromises segmentation stability. We mitigate this issue by repurposing SAM2 as a video polyp segmenter with two training-free modules. In particular, the intra-association filtering module eliminates spatial inaccuracies originating from the detecting stage, reducing false positives. The inter-association refinement module adaptively updates the memory bank to prevent error propagation over time, enhancing temporal coherence. Both modules work synergistically to stabilize SAM2, achieving cutting-edge performance in both in-domain and out-of-domain scenarios. Furthermore, we demonstrate the robust tracking capabilities of FreeVPS in long-untrimmed colonoscopy videos, underscoring its potential reliable clinical analysis.
Authors: Stelios Mylonas, Symeon Papadopoulos
Abstract: The increasing realism and accessibility of deepfakes have raised critical concerns about media authenticity and information integrity. Despite recent advances, deepfake detection models often struggle to generalize beyond their training distributions, particularly when applied to media content found in the wild. In this work, we present a robust video deepfake detection framework with strong generalization that takes advantage of the rich facial representations learned by face foundation models. Our method is built on top of FSFM, a self-supervised model trained on real face data, and is further fine-tuned using an ensemble of deepfake datasets spanning both face-swapping and face-reenactment manipulations. To enhance discriminative power, we incorporate triplet loss variants during training, guiding the model to produce more separable embeddings between real and fake samples. Additionally, we explore attribution-based supervision schemes, where deepfakes are categorized by manipulation type or source dataset, to assess their impact on generalization. Extensive experiments across diverse evaluation benchmarks demonstrate the effectiveness of our approach, especially in challenging real-world scenarios.
Authors: Chenguang Liu, Chisheng Wang, Yuhua Cai, Chuanhua Zhu, Qingquan Li
Abstract: Line segment detection in images has been studied for several decades. Existing line segment detectors can be roughly divided into two categories: generic line segment detectors and wireframe line segment detectors. Generic line segment detectors aim to detect all meaningful line segments in images and traditional approaches usually fall into this category. Recent deep learning based approaches are mostly wireframe line segment detectors. They detect only line segments that are geometrically meaningful and have large spatial support. Due to the difference in the aim of design, the performance of generic line segment detectors for the task of wireframe line segment detection won't be satisfactory, and vice versa. In this work, we propose a robust framework that can be used for both generic line segment detection and wireframe line segment detection. The proposed method is an improved version of the Pixel Orientation Estimation (POE) method. It is thus named as POEv2. POEv2 detects line segments from edge strength maps, and can be combined with any edge detector. We show in our experiments that by combining the proposed POEv2 with an efficient edge detector, it achieves state-of-the-art performance on three publicly available datasets.
Authors: Qiyao Xu, Qiming Wu, Xiaowei Li
Abstract: Segment Anything Model (SAM) has demonstrated remarkable capabilities in solving light field salient object detection (LF SOD). However, most existing models tend to neglect the extraction of prompt information under this task. Meanwhile, traditional models ignore the analysis of frequency-domain information, which leads to small objects being overwhelmed by noise. In this paper, we put forward a novel model called self-prompting light field segment anything model (SPLF-SAM), equipped with unified multi-scale feature embedding block (UMFEB) and a multi-scale adaptive filtering adapter (MAFA). UMFEB is capable of identifying multiple objects of varying sizes, while MAFA, by learning frequency features, effectively prevents small objects from being overwhelmed by noise. Extensive experiments have demonstrated the superiority of our method over ten state-of-the-art (SOTA) LF SOD methods. Our code will be available at https://github.com/XucherCH/splfsam.
Authors: Yue Wu, Yufan Wu, Wen Li, Yuxi Lu, Kairui Feng, Xuanhong Chen
Abstract: Despite significant progress in 3D avatar reconstruction, it still faces challenges such as high time complexity, sensitivity to data quality, and low data utilization. We propose FastAvatar, a feedforward 3D avatar framework capable of flexibly leveraging diverse daily recordings (e.g., a single image, multi-view observations, or monocular video) to reconstruct a high-quality 3D Gaussian Splatting (3DGS) model within seconds, using only a single unified model. FastAvatar's core is a Large Gaussian Reconstruction Transformer featuring three key designs: First, a variant VGGT-style transformer architecture aggregating multi-frame cues while injecting initial 3D prompt to predict an aggregatable canonical 3DGS representation; Second, multi-granular guidance encoding (camera pose, FLAME expression, head pose) mitigating animation-induced misalignment for variable-length inputs; Third, incremental Gaussian aggregation via landmark tracking and sliced fusion losses. Integrating these features, FastAvatar enables incremental reconstruction, i.e., improving quality with more observations, unlike prior work wasting input data. This yields a quality-speed-tunable paradigm for highly usable avatar modeling. Extensive experiments show that FastAvatar has higher quality and highly competitive speed compared to existing methods.
Authors: Ahmed Emam, Mohamed Elbassiouny, Julius Miller, Patrick Donworth, Sabine Seidel, Ribana Roscher
Abstract: Pollinator insects such as honeybees and bumblebees are vital to global food production and ecosystem stability, yet their populations are declining due to increasing anthropogenic and environmental stressors. To support scalable, automated pollinator monitoring, we introduce BuzzSet, a new large-scale dataset of high-resolution pollinator images collected in real agricultural field conditions. BuzzSet contains 7856 manually verified and labeled images, with over 8000 annotated instances across three classes: honeybees, bumblebees, and unidentified insects. Initial annotations were generated using a YOLOv12 model trained on external data and refined via human verification using open-source labeling tools. All images were preprocessed into 256~$\times$~256 tiles to improve the detection of small insects. We provide strong baselines using the RF-DETR transformer-based object detector. The model achieves high F1-scores of 0.94 and 0.92 for honeybee and bumblebee classes, respectively, with confusion matrix results showing minimal misclassification between these categories. The unidentified class remains more challenging due to label ambiguity and lower sample frequency, yet still contributes useful insights for robustness evaluation. Overall detection quality is strong, with a best mAP@0.50 of 0.559. BuzzSet offers a valuable benchmark for small object detection, class separation under label noise, and ecological computer vision.
Authors: Shu Shen, C. L. Philip Chen, Tong Zhang
Abstract: Multimodal learning has significantly enhanced machine learning performance but still faces numerous challenges and limitations. Imbalanced multimodal learning is one of the problems extensively studied in recent works and is typically mitigated by modulating the learning of each modality. However, we find that these methods typically hinder the dominant modality's learning to promote weaker modalities, which affects overall multimodal performance. We analyze the cause of this issue and highlight a commonly overlooked problem: optimization bias within networks. To address this, we propose Adaptive Intra-Network Modulation (AIM) to improve balanced modality learning. AIM accounts for differences in optimization state across parameters and depths within the network during modulation, achieving balanced multimodal learning without hindering either dominant or weak modalities for the first time. Specifically, AIM decouples the dominant modality's under-optimized parameters into Auxiliary Blocks and encourages reliance on these performance-degraded blocks for joint training with weaker modalities. This approach effectively prevents suppression of weaker modalities while enabling targeted optimization of under-optimized parameters to improve the dominant modality. Additionally, AIM assesses modality imbalance level across network depths and adaptively adjusts modulation strength at each depth. Experimental results demonstrate that AIM outperforms state-of-the-art imbalanced modality learning methods across multiple benchmarks and exhibits strong generalizability across different backbones, fusion strategies, and optimizers.
Authors: Jakob Seitz, Tobias Lengfeld, Radu Timofte
Abstract: Handwritten Mathematical Expression Recognition is foundational for educational technologies, enabling applications like digital note-taking and automated grading. While modern encoder-decoder architectures with large language models excel at LaTeX generation, they lack explicit symbol-to-trace alignment, a critical limitation for error analysis, interpretability, and spatially aware interactive applications requiring selective content updates. This paper introduces a structural recognition approach with two innovations: 1 an automatic annotation system that uses a neural network to map LaTeX equations to raw traces, automatically generating annotations for symbol segmentation, classification, and spatial relations, and 2 a modular structural recognition system that independently optimizes segmentation, classification, and relation prediction. By leveraging a dataset enriched with structural annotations from our auto-labeling system, the proposed recognition system combines graph-based trace sorting, a hybrid convolutional-recurrent network, and transformer-based correction to achieve competitive performance on the CROHME-2023 benchmark. Crucially, our structural recognition system generates a complete graph structure that directly links handwritten traces to predicted symbols, enabling transparent error analysis and interpretable outputs.
Authors: Han Jiao, Jiakai Sun, Yexing Xu, Lei Zhao, Wei Xing, Huaizhong Lin
Abstract: 3D Gaussian Splatting, known for enabling high-quality static scene reconstruction with fast rendering, is increasingly being applied to dynamic scene reconstruction. A common strategy involves learning a deformation field to model the temporal changes of a canonical set of 3D Gaussians. However, these deformation-based methods often produce blurred renderings and lose fine motion details in highly dynamic regions due to the inherent limitations of a single, unified model in representing diverse motion patterns. To address these challenges, we introduce Motion-Aware Partitioning of Deformable 3D Gaussian Splatting (MAPo), a novel framework for high-fidelity dynamic scene reconstruction. Its core is a dynamic score-based partitioning strategy that distinguishes between high- and low-dynamic 3D Gaussians. For high-dynamic 3D Gaussians, we recursively partition them temporally and duplicate their deformation networks for each new temporal segment, enabling specialized modeling to capture intricate motion details. Concurrently, low-dynamic 3DGs are treated as static to reduce computational costs. However, this temporal partitioning strategy for high-dynamic 3DGs can introduce visual discontinuities across frames at the partition boundaries. To address this, we introduce a cross-frame consistency loss, which not only ensures visual continuity but also further enhances rendering quality. Extensive experiments demonstrate that MAPo achieves superior rendering quality compared to baselines while maintaining comparable computational costs, particularly in regions with complex or rapid motions.
Authors: Xiuchao Wu, Pengfei Zhu, Jiangjing Lyu, Xinguo Liu, Jie Guo, Yanwen Guo, Weiwei Xu, Chengfei Lyu
Abstract: Recovering material information from images has been extensively studied in computer graphics and vision. Recent works in material estimation leverage diffusion model showing promising results. However, these diffusion-based methods adopt a multi-step denoising strategy, which is time-consuming for each estimation. Such stochastic inference also conflicts with the deterministic material estimation task, leading to a high variance estimated results. In this paper, we introduce StableIntrinsic, a one-step diffusion model for multi-view material estimation that can produce high-quality material parameters with low variance. To address the overly-smoothing problem in one-step diffusion, StableIntrinsic applies losses in pixel space, with each loss designed based on the properties of the material. Additionally, StableIntrinsic introduces a Detail Injection Network (DIN) to eliminate the detail loss caused by VAE encoding, while further enhancing the sharpness of material prediction results. The experimental results indicate that our method surpasses the current state-of-the-art techniques by achieving a $9.9\%$ improvement in the Peak Signal-to-Noise Ratio (PSNR) of albedo, and by reducing the Mean Square Error (MSE) for metallic and roughness by $44.4\%$ and $60.0\%$, respectively.
Authors: Shay Shomer Chai, Wenxuan Peng, Bharath Hariharan, Hadar Averbuch-Elor
Abstract: Text-to-image generation has recently seen remarkable success, granting users with the ability to create high-quality images through the use of text. However, contemporary methods face challenges in capturing the precise semantics conveyed by complex multi-object prompts. Consequently, many works have sought to mitigate such semantic misalignments, typically via inference-time schemes that modify the attention layers of the denoising networks. However, prior work has mostly utilized coarse metrics, such as the cosine similarity between text and image CLIP embeddings, or human evaluations, which are challenging to conduct on a larger-scale. In this work, we perform a case study on colors -- a fundamental attribute commonly associated with objects in text prompts, which offer a rich test bed for rigorous evaluation. Our analysis reveals that pretrained models struggle to generate images that faithfully reflect multiple color attributes-far more so than with single-color prompts-and that neither inference-time techniques nor existing editing methods reliably resolve these semantic misalignments. Accordingly, we introduce a dedicated image editing technique, mitigating the issue of multi-object semantic alignment for prompts containing multiple colors. We demonstrate that our approach significantly boosts performance over a wide range of metrics, considering images generated by various text-to-image diffusion-based techniques.
Authors: Muhammad Ali, Omar Ali AlSuwaidi
Abstract: In the realm of waste management, automating the sorting process for non-biodegradable materials presents considerable challenges due to the complexity and variability of waste streams. To address these challenges, we introduce an enhanced neural architecture that builds upon an existing Encoder-Decoder structure to improve the accuracy and efficiency of waste sorting systems. Our model integrates several key innovations: a Comprehensive Attention Block within the decoder, which refines feature representations by combining convolutional and upsampling operations. In parallel, we utilize attention through the Mamba architecture, providing an additional performance boost. We also introduce a Data Fusion Block that fuses images with more than three channels. To achieve this, we apply PCA transformation to reduce the dimensionality while retaining the maximum variance and essential information across three dimensions, which are then used for further processing. We evaluated the model on RGB, hyperspectral, multispectral, and a combination of RGB and hyperspectral data. The results demonstrate that our approach outperforms existing methods by a significant margin.
Authors: Christian Marzahl, Brian Napora
Abstract: Mitotic figure (MF) detection in histopathology images is challenging due to large variations in slide scanners, staining protocols, tissue types, and the presence of artifacts. This paper presents a collection of training techniques - a bag of tricks - that enable robust, real-time MF detection across diverse domains. We build on the efficient RTMDet single stage object detector to achieve high inference speed suitable for clinical deployment. Our method addresses scanner variability and tumor heterogeneity via extensive multi-domain training data, balanced sampling, and careful augmentation. Additionally, we employ targeted, hard negative mining on necrotic and debris tissue to reduce false positives. In a grouped 5-fold cross-validation across multiple MF datasets, our model achieves an F1 score between 0.78 and 0.84. On the preliminary test set of the MItosis DOmain Generalization (MIDOG) 2025 challenge, our single-stage RTMDet-S based approach reaches an F1 of 0.81, outperforming larger models and demonstrating adaptability to new, unfamiliar domains. The proposed solution offers a practical trade-off between accuracy and speed, making it attractive for real-world clinical adoption.
Authors: Shenqi Wang, Guangzhi Tang
Abstract: Event-based camera has emerged as a promising paradigm for robot perception, offering advantages with high temporal resolution, high dynamic range, and robustness to motion blur. However, existing deep learning-based event processing methods often fail to fully leverage the sparse nature of event data, complicating their integration into resource-constrained edge applications. While neuromorphic computing provides an energy-efficient alternative, spiking neural networks struggle to match of performance of state-of-the-art models in complex event-based vision tasks, like object detection and optical flow. Moreover, achieving high activation sparsity in neural networks is still difficult and often demands careful manual tuning of sparsity-inducing loss terms. Here, we propose Context-aware Sparse Spatiotemporal Learning (CSSL), a novel framework that introduces context-aware thresholding to dynamically regulate neuron activations based on the input distribution, naturally reducing activation density without explicit sparsity constraints. Applied to event-based object detection and optical flow estimation, CSSL achieves comparable or superior performance to state-of-the-art methods while maintaining extremely high neuronal sparsity. Our experimental results highlight CSSL's crucial role in enabling efficient event-based vision for neuromorphic processing.
Authors: Kaixuan Lu, Mehmet Onurcan Kaya, Dim P. Papadopoulos
Abstract: Video Instance Segmentation (VIS) faces significant annotation challenges due to its dual requirements of pixel-level masks and temporal consistency labels. While recent unsupervised methods like VideoCutLER eliminate optical flow dependencies through synthetic data, they remain constrained by the synthetic-to-real domain gap. We present AutoQ-VIS, a novel unsupervised framework that bridges this gap through quality-guided self-training. Our approach establishes a closed-loop system between pseudo-label generation and automatic quality assessment, enabling progressive adaptation from synthetic to real videos. Experiments demonstrate state-of-the-art performance with 52.6 $\text{AP}_{50}$ on YouTubeVIS-2019 val set, surpassing the previous state-of-the-art VideoCutLER by 4.4$\%$, while requiring no human annotations. This demonstrates the viability of quality-aware self-training for unsupervised VIS. The source code of our method is available at https://github.com/wcbup/AutoQ-VIS.
Authors: Linkuan Zhou, Zhexin Chen, Yufei Shen, Junlin Xu, Ping Xuan, Yixin Zhu, Yuqi Fang, Cong Cong, Leyi Wei, Ran Su, Jia Zhou, Qiangguo Jin
Abstract: Automated segmentation of the fetal head in ultrasound images is critical for prenatal monitoring. However, achieving robust segmentation remains challenging due to the poor quality of ultrasound images and the lack of annotated data. Semi-supervised methods alleviate the lack of annotated data but struggle with the unique characteristics of fetal head ultrasound images, making it challenging to generate reliable pseudo-labels and enforce effective consistency regularization constraints. To address this issue, we propose a novel semi-supervised framework, ERSR, for fetal head ultrasound segmentation. Our framework consists of the dual-scoring adaptive filtering strategy, the ellipse-constrained pseudo-label refinement, and the symmetry-based multiple consistency regularization. The dual-scoring adaptive filtering strategy uses boundary consistency and contour regularity criteria to evaluate and filter teacher outputs. The ellipse-constrained pseudo-label refinement refines these filtered outputs by fitting least-squares ellipses, which strengthens pixels near the center of the fitted ellipse and suppresses noise simultaneously. The symmetry-based multiple consistency regularization enforces multi-level consistency across perturbed images, symmetric regions, and between original predictions and pseudo-labels, enabling the model to capture robust and stable shape representations. Our method achieves state-of-the-art performance on two benchmarks. On the HC18 dataset, it reaches Dice scores of 92.05% and 95.36% with 10% and 20% labeled data, respectively. On the PSFH dataset, the scores are 91.68% and 93.70% under the same settings.
Authors: Yilin Zhang, Cai Xu, You Wu, Ziyu Guan, Wei Zhao
Abstract: Deep neural networks often produce overconfident predictions, undermining their reliability in safety-critical applications. This miscalibration is further exacerbated under distribution shift, where test data deviates from the training distribution due to environmental or acquisition changes. While existing approaches improve calibration through training-time regularization or post-hoc adjustment, their reliance on access to or simulation of target domains limits their practicality in real-world scenarios. In this paper, we propose a novel calibration framework that operates without access to target domain information. From a frequency-domain perspective, we identify that distribution shifts often distort high-frequency visual cues exploited by deep models, and introduce a low-frequency filtering strategy to encourage reliance on domain-invariant features. However, such information loss may degrade In-Distribution (ID) calibration performance. Therefore, we further propose a gradient-based rectification mechanism that enforces ID calibration as a hard constraint during optimization. Experiments on synthetic and real-world shifted datasets, including CIFAR-10/100-C and WILDS, demonstrate that our method significantly improves calibration under distribution shift while maintaining strong in-distribution performance.
Authors: Xiaoqi Wang, Yun Zhang, Weisi Lin
Abstract: Machine vision systems (MVS) are intrinsically vulnerable to performance degradation under adverse visual conditions. To address this, we propose a machine-centric image quality assessment (MIQA) framework that quantifies the impact of image degradations on MVS performance. We establish an MIQA paradigm encompassing the end-to-end assessment workflow. To support this, we construct a machine-centric image quality database (MIQD-2.5M), comprising 2.5 million samples that capture distinctive degradation responses in both consistency and accuracy metrics, spanning 75 vision models, 250 degradation types, and three representative vision tasks. We further propose a region-aware MIQA (RA-MIQA) model to evaluate MVS visual quality through fine-grained spatial degradation analysis. Extensive experiments benchmark the proposed RA-MIQA against seven human visual system (HVS)-based IQA metrics and five retrained classical backbones. Results demonstrate RA-MIQA's superior performance in multiple dimensions, e.g., achieving SRCC gains of 13.56% on consistency and 13.37% on accuracy for image classification, while also revealing task-specific degradation sensitivities. Critically, HVS-based metrics prove inadequate for MVS quality prediction, while even specialized MIQA models struggle with background degradations, accuracy-oriented estimation, and subtle distortions. This study can advance MVS reliability and establish foundations for machine-centric image processing and optimization. The model and code are available at: https://github.com/XiaoqiWang/MIQA.
Authors: Binjie Zhang, Mike Zheng Shou
Abstract: In egocentric scenarios, anticipating both the next action and its visual outcome is essential for understanding human-object interactions and for enabling robotic planning. However, existing paradigms fall short of jointly modeling these aspects. Vision-Language-Action (VLA) models focus on action prediction but lack explicit modeling of how actions influence the visual scene, while video prediction models generate future frames without conditioning on specific actions, often resulting in implausible or contextually inconsistent outcomes. To bridge this gap, we propose a unified two-stage predictive framework that jointly models action and visual future in egocentric scenarios, conditioned on hand trajectories. In the first stage, we perform consecutive state modeling to process heterogeneous inputs (visual observations, language, and action history) and explicitly predict future hand trajectories. In the second stage, we introduce causal cross-attention to fuse multi-modal cues, leveraging inferred action signals to guide an image-based Latent Diffusion Model (LDM) for frame-by-frame future video generation. Our approach is the first unified model designed to handle both egocentric human activity understanding and robotic manipulation tasks, providing explicit predictions of both upcoming actions and their visual consequences. Extensive experiments on Ego4D, BridgeData, and RLBench demonstrate that our method outperforms state-of-the-art baselines in both action prediction and future video synthesis.
Authors: Long Chen, Ashiv Patel, Mengyun Qiao, Mohammad Yousuf Salmasi, Salah A. Hammouche, Vasilis Stavrinides, Jasleen Nagi, Soodeh Kalaie, Xiao Yun Xu, Wenjia Bai, Declan P. O'Regan
Abstract: Personalized, accurate prediction of aortic aneurysm progression is essential for timely intervention but remains challenging due to the need to model both subtle local deformations and global anatomical changes within complex 3D geometries. We propose MCMeshGAN, the first multimodal conditional mesh-to-mesh generative adversarial network for 3D aneurysm growth prediction. MCMeshGAN introduces a dual-branch architecture combining a novel local KNN-based convolutional network (KCN) to preserve fine-grained geometric details and a global graph convolutional network (GCN) to capture long-range structural context, overcoming the over-smoothing limitations of deep GCNs. A dedicated condition branch encodes clinical attributes (age, sex) and the target time interval to generate anatomically plausible, temporally controlled predictions, enabling retrospective and prospective modeling. We curated TAAMesh, a new longitudinal thoracic aortic aneurysm mesh dataset consisting of 590 multimodal records (CT scans, 3D meshes, and clinical data) from 208 patients. Extensive experiments demonstrate that MCMeshGAN consistently outperforms state-of-the-art baselines in both geometric accuracy and clinically important diameter estimation. This framework offers a robust step toward clinically deployable, personalized 3D disease trajectory modeling. The source code for MCMeshGAN and the baseline methods is publicly available at https://github.com/ImperialCollegeLondon/MCMeshGAN.
Authors: Oussama Hadjerci, Antoine Letienne, Mohamed Abbas Hedjazi, Adel Hafiane
Abstract: Self-supervised learning (SSL) has emerged as a powerful technique for learning visual representations. While recent SSL approaches achieve strong results in global image understanding, they are limited in capturing the structured representation in scenes. In this work, we propose a self-supervised approach that progressively builds structured visual representations by combining semantic grouping, instance level separation, and hierarchical structuring. Our approach, based on a novel ProtoScale module, captures visual elements across multiple spatial scales. Unlike common strategies like DINO that rely on random cropping and global embeddings, we preserve full scene context across augmented views to improve performance in dense prediction tasks. We validate our method on downstream object detection tasks using a combined subset of multiple datasets (COCO and UA-DETRAC). Experimental results show that our method learns object centric representations that enhance supervised object detection and outperform the state-of-the-art methods, even when trained with limited annotated data and fewer fine-tuning epochs.
Authors: Fran\c{c}ois G. Landry, Moulay A. Akhloufi
Abstract: With the introduction of vehicles with autonomous capabilities on public roads, predicting pedestrian crossing intention has emerged as an active area of research. The task of predicting pedestrian crossing intention involves determining whether pedestrians in the scene are likely to cross the road or not. In this work, we propose TrajFusionNet, a novel transformer-based model that combines future pedestrian trajectory and vehicle speed predictions as priors for predicting crossing intention. TrajFusionNet comprises two branches: a Sequence Attention Module (SAM) and a Visual Attention Module (VAM). The SAM branch learns from a sequential representation of the observed and predicted pedestrian trajectory and vehicle speed. Complementarily, the VAM branch enables learning from a visual representation of the predicted pedestrian trajectory by overlaying predicted pedestrian bounding boxes onto scene images. By utilizing a small number of lightweight modalities, TrajFusionNet achieves the lowest total inference time (including model runtime and data preprocessing) among current state-of-the-art approaches. In terms of performance, it achieves state-of-the-art results across the three most commonly used datasets for pedestrian crossing intention prediction.
Authors: Hui Zhang, Jianghui Cai, Haifeng Yang, Ali Luo, Yuqing Yang, Xiao Kong, Zhichao Ding, Lichan Zhou, Qin Han
Abstract: Sky background subtraction is a critical step in Multi-objective Fiber spectra process. However, current subtraction relies mainly on sky fiber spectra to build Super Sky. These average spectra are lacking in the modeling of the environment surrounding the objects. To address this issue, a sky background estimation model: Sky background building based on Mutual Information (SMI) is proposed. SMI based on mutual information and incremental training approach. It utilizes spectra from all fibers in the plate to estimate the sky background. SMI contains two main networks, the first network applies a wavelength calibration module to extract sky features from spectra, and can effectively solve the feature shift problem according to the corresponding emission position. The second network employs an incremental training approach to maximize mutual information between representations of different spectra to capturing the common component. Then, it minimizes the mutual information between adjoining spectra representations to obtain individual components. This network yields an individual sky background at each location of the object. To verify the effectiveness of the method in this paper, we conducted experiments on the spectra of LAMOST. Results show that SMI can obtain a better object sky background during the observation, especially in the blue end.
Authors: Narges Takhtkeshha, Gabriele Mazzacca, Fabio Remondino, Juha Hyypp\"a, Gottfried Mandlburger
Abstract: Monitoring urban tree dynamics is vital for supporting greening policies and reducing risks to electrical infrastructure. Airborne laser scanning has advanced large-scale tree management, but challenges remain due to complex urban environments and tree variability. Multispectral (MS) light detection and ranging (LiDAR) improves this by capturing both 3D spatial and spectral data, enabling detailed mapping. This study explores tree point extraction using MS-LiDAR and deep learning (DL) models. Three state-of-the-art models are evaluated: Superpoint Transformer (SPT), Point Transformer V3 (PTv3), and Point Transformer V1 (PTv1). Results show the notable time efficiency and accuracy of SPT, with a mean intersection over union (mIoU) of 85.28%. The highest detection accuracy is achieved by incorporating pseudo normalized difference vegetation index (pNDVI) with spatial data, reducing error rate by 10.61 percentage points (pp) compared to using spatial information alone. These findings highlight the potential of MS-LiDAR and DL to improve tree extraction and further tree inventories.
Authors: Ziyun Qian, Runyu Xiao, Shuyuan Tu, Wei Xue, Dingkang Yang, Mingcheng Li, Dongliang Kou, Minghao Han, Zizhi Chen, Lihua Zhang
Abstract: Recent advances in motion generation show remarkable progress. However, several limitations remain: (1) Existing pose-guided character motion transfer methods merely replicate motion without learning its style characteristics, resulting in inexpressive characters. (2) Motion style transfer methods rely heavily on motion capture data, which is difficult to obtain. (3) Generated motions sometimes violate physical laws. To address these challenges, this paper pioneers a new task: Video-to-Video Motion Personalization. We propose a novel framework, PersonaAnimator, which learns personalized motion patterns directly from unconstrained videos. This enables personalized motion transfer. To support this task, we introduce PersonaVid, the first video-based personalized motion dataset. It contains 20 motion content categories and 120 motion style categories. We further propose a Physics-aware Motion Style Regularization mechanism to enforce physical plausibility in the generated motions. Extensive experiments show that PersonaAnimator outperforms state-of-the-art motion transfer methods and sets a new benchmark for the Video-to-Video Motion Personalization task.
Authors: Imad Ali Shah, Jiarong Li, Roshan George, Tim Brophy, Enda Ward, Martin Glavin, Edward Jones, Brian Deegan
Abstract: Hyperspectral imaging (HSI) offers a transformative sensing modality for Advanced Driver Assistance Systems (ADAS) and autonomous driving (AD) applications, enabling material-level scene understanding through fine spectral resolution beyond the capabilities of traditional RGB imaging. This paper presents the first comprehensive review of HSI for automotive applications, examining the strengths, limitations, and suitability of current HSI technologies in the context of ADAS/AD. In addition to this qualitative review, we analyze 216 commercially available HSI and multispectral imaging cameras, benchmarking them against key automotive criteria: frame rate, spatial resolution, spectral dimensionality, and compliance with AEC-Q100 temperature standards. Our analysis reveals a significant gap between HSI's demonstrated research potential and its commercial readiness. Only four cameras meet the defined performance thresholds, and none comply with AEC-Q100 requirements. In addition, the paper reviews recent HSI datasets and applications, including semantic segmentation for road surface classification, pedestrian separability, and adverse weather perception. Our review shows that current HSI datasets are limited in terms of scale, spectral consistency, the number of spectral channels, and environmental diversity, posing challenges for the development of perception algorithms and the adequate validation of HSI's true potential in ADAS/AD applications. This review paper establishes the current state of HSI in automotive contexts as of 2025 and outlines key research directions toward practical integration of spectral imaging in ADAS and autonomous systems.
Authors: Moussa Kassem Sbeyti, Nadja Klein, Michelle Karg, Christian Wirth, Sahin Albayrak
Abstract: Active learning (AL) for real-world object detection faces computational and reliability challenges that limit practical deployment. Developing new AL methods requires training multiple detectors across iterations to compare against existing approaches. This creates high costs for autonomous driving datasets where the training of one detector requires up to 282 GPU hours. Additionally, AL method rankings vary substantially across validation sets, compromising reliability in safety-critical transportation systems. We introduce object-based set similarity ($\mathrm{OSS}$), a metric that addresses these challenges. $\mathrm{OSS}$ (1) quantifies AL method effectiveness without requiring detector training by measuring similarity between training sets and target domains using object-level features. This enables the elimination of ineffective AL methods before training. Furthermore, $\mathrm{OSS}$ (2) enables the selection of representative validation sets for robust evaluation. We validate our similarity-based approach on three autonomous driving datasets (KITTI, BDD100K, CODA) using uncertainty-based AL methods as a case study with two detector architectures (EfficientDet, YOLOv3). This work is the first to unify AL training and evaluation strategies in object detection based on object similarity. $\mathrm{OSS}$ is detector-agnostic, requires only labeled object crops, and integrates with existing AL pipelines. This provides a practical framework for deploying AL in real-world applications where computational efficiency and evaluation reliability are critical. Code is available at https://mos-ks.github.io/publications/.
Authors: Lechun You, Zhonghua Wu, Weide Liu, Xulei Yang, Jun Cheng, Wei Zhou, Bharadwaj Veeravalli, Guosheng Lin
Abstract: Current methods for 3D semantic segmentation propose training models with limited annotations to address the difficulty of annotating large, irregular, and unordered 3D point cloud data. They usually focus on the 3D domain only, without leveraging the complementary nature of 2D and 3D data. Besides, some methods extend original labels or generate pseudo labels to guide the training, but they often fail to fully use these labels or address the noise within them. Meanwhile, the emergence of comprehensive and adaptable foundation models has offered effective solutions for segmenting 2D data. Leveraging this advancement, we present a novel approach that maximizes the utility of sparsely available 3D annotations by incorporating segmentation masks generated by 2D foundation models. We further propagate the 2D segmentation masks into the 3D space by establishing geometric correspondences between 3D scenes and 2D views. We extend the highly sparse annotations to encompass the areas delineated by 3D masks, thereby substantially augmenting the pool of available labels. Furthermore, we apply confidence- and uncertainty-based consistency regularization on augmentations of the 3D point cloud and select the reliable pseudo labels, which are further spread on the 3D masks to generate more labels. This innovative strategy bridges the gap between limited 3D annotations and the powerful capabilities of 2D foundation models, ultimately improving the performance of 3D weakly supervised segmentation.
Authors: Fayaz Ali, Muhammad Zawish, Steven Davy, Radu Timofte
Abstract: Transformers have demonstrated promising performance in computer vision tasks, including image super-resolution (SR). The quadratic computational complexity of window self-attention mechanisms in many transformer-based SR methods forces the use of small, fixed windows, limiting the receptive field. In this paper, we propose a new approach by embedding the wavelet transform within a hierarchical transformer framework, called (WaveHiT-SR). First, using adaptive hierarchical windows instead of static small windows allows to capture features across different levels and greatly improve the ability to model long-range dependencies. Secondly, the proposed model utilizes wavelet transforms to decompose images into multiple frequency subbands, allowing the network to focus on both global and local features while preserving structural details. By progressively reconstructing high-resolution images through hierarchical processing, the network reduces computational complexity without sacrificing performance. The multi-level decomposition strategy enables the network to capture fine-grained information in lowfrequency components while enhancing high-frequency textures. Through extensive experimentation, we confirm the effectiveness and efficiency of our WaveHiT-SR. Our refined versions of SwinIR-Light, SwinIR-NG, and SRFormer-Light deliver cutting-edge SR results, achieving higher efficiency with fewer parameters, lower FLOPs, and faster speeds.
Authors: Taebaek Hwang, Minseo Kim, Gisang Lee, Seonuk Kim, Hyunjun Eun
Abstract: Understanding and reasoning over text within visual contexts poses a significant challenge for Vision-Language Models (VLMs), given the complexity and diversity of real-world scenarios. To address this challenge, text-rich Visual Question Answering (VQA) datasets and benchmarks have emerged for high-resource languages like English. However, a critical gap persists for low-resource languages such as Korean, where the lack of comprehensive benchmarks hinders robust model evaluation and comparison. To bridge this gap, we introduce KRETA, a benchmark for Korean Reading and rEasoning in Text-rich VQA Attuned to diverse visual contexts. KRETA facilitates an in-depth evaluation of both visual text understanding and reasoning capabilities, while also supporting a multifaceted assessment across 15 domains and 26 image types. Additionally, we introduce a semi-automated VQA generation pipeline specifically optimized for text-rich settings, leveraging refined stepwise image decomposition and a rigorous seven-metric evaluation protocol to ensure data quality. While KRETA is tailored for Korean, we hope our adaptable and extensible pipeline will facilitate the development of similar benchmarks in other languages, thereby accelerating multilingual VLM research. The code and dataset for KRETA are available at https://github.com/tabtoyou/KRETA.
Authors: Gianluca Guzzetta
Abstract: In this paper, we present a comprehensive study and analysis of the Chan-Vese algorithm for image segmentation. We employ a discretized scheme derived from the empirical study of the Chan-Vese model's functional energy and its partial differential equation based on its level set function. We provide a proof of the results and an implementation using MATLAB. Leveraging modern computer vision methodologies, we propose a functional segmentation loss based on active contours, utilizing pytorch.nn.ModuleLoss and a level set based on the Chan-Vese algorithm. We compare our results with common computer vision segmentation datasets and evaluate the performance of classical loss functions against our proposed method. All code and materials used are available at https://github.com/gguzzy/chan_vese_functional_loss.
Authors: Oliver Grainge, Sania Waheed, Jack Stilgoe, Michael Milford, Shoaib Ehsan
Abstract: Geo-localization is the task of identifying the location of an image using visual cues alone. It has beneficial applications, such as improving disaster response, enhancing navigation, and geography education. Recently, Vision-Language Models (VLMs) are increasingly demonstrating capabilities as accurate image geo-locators. This brings significant privacy risks, including those related to stalking and surveillance, considering the widespread uses of AI models and sharing of photos on social media. The precision of these models is likely to improve in the future. Despite these risks, there is little work on systematically evaluating the geolocation precision of Generative VLMs, their limits and potential for unintended inferences. To bridge this gap, we conduct a comprehensive assessment of the geolocation capabilities of 25 state-of-the-art VLMs on four benchmark image datasets captured in diverse environments. Our results offer insight into the internal reasoning of VLMs and highlight their strengths, limitations, and potential societal risks. Our findings indicate that current VLMs perform poorly on generic street-level images yet achieve notably high accuracy (61\%) on images resembling social media content, raising significant and urgent privacy concerns.
Authors: Seongheon Park, Yixuan Li
Abstract: Object hallucination in large vision-language models presents a significant challenge to their safe deployment in real-world applications. Recent works have proposed object-level hallucination scores to estimate the likelihood of object hallucination; however, these methods typically adopt either a global or local perspective in isolation, which may limit detection reliability. In this paper, we introduce GLSim, a novel training-free object hallucination detection framework that leverages complementary global and local embedding similarity signals between image and text modalities, enabling more accurate and reliable hallucination detection in diverse scenarios. We comprehensively benchmark existing object hallucination detection methods and demonstrate that GLSim achieves superior detection performance, outperforming competitive baselines by a significant margin.
Authors: Yuhao Chen, Shubin Chen, Liang Lin, Guangrun Wang
Abstract: Language-driven image segmentation is a fundamental task in vision-language understanding, requiring models to segment regions of an image corresponding to natural language expressions. Traditional methods approach this as a discriminative problem, assigning each pixel to foreground or background based on semantic alignment. Recently, diffusion models have been introduced to this domain, but existing approaches remain image-centric: they either (i) use image diffusion models as visual feature extractors, (ii) synthesize segmentation data via image generation to train discriminative models, or (iii) perform diffusion inversion to extract attention cues from pre-trained image diffusion models-thereby treating segmentation as an auxiliary process. In this paper, we propose GS (Generative Segmentation), a novel framework that formulates segmentation itself as a generative task via label diffusion. Instead of generating images conditioned on label maps and text, GS reverses the generative process: it directly generates segmentation masks from noise, conditioned on both the input image and the accompanying language description. This paradigm makes label generation the primary modeling target, enabling end-to-end training with explicit control over spatial and semantic fidelity. To demonstrate the effectiveness of our approach, we evaluate GS on Panoptic Narrative Grounding (PNG), a representative and challenging benchmark for multimodal segmentation that requires panoptic-level reasoning guided by narrative captions. Experimental results show that GS significantly outperforms existing discriminative and diffusion-based methods, setting a new state-of-the-art for language-driven segmentation.
Authors: Manogna Sreenivas, Soma Biswas
Abstract: In dynamic environments, unfamiliar objects and distribution shifts are often encountered, which challenge the generalization abilities of the deployed trained models. This work addresses Incremental Test Time Adaptation of Vision Language Models, tackling scenarios where unseen classes and unseen domains continuously appear during testing. Unlike traditional Test Time Adaptation approaches, where the test stream comes only from a predefined set of classes, our framework allows models to adapt simultaneously to both covariate and label shifts, actively incorporating new classes as they emerge. Towards this goal, we establish a new benchmark for ITTA, integrating single image TTA methods for VLMs with active labeling techniques that query an oracle for samples potentially representing unseen classes during test time. We propose a segmentation assisted active labeling module, termed SegAssist, which is training free and repurposes the segmentation capabilities of VLMs to refine active sample selection, prioritizing samples likely to belong to unseen classes. Extensive experiments on several benchmark datasets demonstrate the potential of SegAssist to enhance the performance of VLMs in real world scenarios, where continuous adaptation to emerging data is essential. Project-page:https://manogna-s.github.io/segassist/
Authors: Peng-Hao Hsu, Ke Zhang, Fu-En Wang, Tao Tu, Ming-Feng Li, Yu-Lun Liu, Albert Y. C. Chen, Min Sun, Cheng-Hao Kuo
Abstract: Open-vocabulary (OV) 3D object detection is an emerging field, yet its exploration through image-based methods remains limited compared to 3D point cloud-based methods. We introduce OpenM3D, a novel open-vocabulary multi-view indoor 3D object detector trained without human annotations. In particular, OpenM3D is a single-stage detector adapting the 2D-induced voxel features from the ImGeoNet model. To support OV, it is jointly trained with a class-agnostic 3D localization loss requiring high-quality 3D pseudo boxes and a voxel-semantic alignment loss requiring diverse pre-trained CLIP features. We follow the training setting of OV-3DET where posed RGB-D images are given but no human annotations of 3D boxes or classes are available. We propose a 3D Pseudo Box Generation method using a graph embedding technique that combines 2D segments into coherent 3D structures. Our pseudo-boxes achieve higher precision and recall than other methods, including the method proposed in OV-3DET. We further sample diverse CLIP features from 2D segments associated with each coherent 3D structure to align with the corresponding voxel feature. The key to training a highly accurate single-stage detector requires both losses to be learned toward high-quality targets. At inference, OpenM3D, a highly efficient detector, requires only multi-view images for input and demonstrates superior accuracy and speed (0.3 sec. per scene) on ScanNet200 and ARKitScenes indoor benchmarks compared to existing methods. We outperform a strong two-stage method that leverages our class-agnostic detector with a ViT CLIP-based OV classifier and a baseline incorporating multi-view depth estimator on both accuracy and speed.
Authors: Philippe Zhang, Weili Jiang, Yihao Li, Jing Zhang, Sarah Matta, Yubo Tan, Hui Lin, Haoshen Wang, Jiangtian Pan, Hui Xu, Laurent Borderie, Alexandre Le Guilcher, B\'eatrice Cochener, Chubin Ou, Gwenol\'e Quellec, Mathieu Lamard
Abstract: Age-related Macular Degeneration (AMD) is a prevalent eye condition affecting visual acuity. Anti-vascular endothelial growth factor (anti-VEGF) treatments have been effective in slowing the progression of neovascular AMD, with better outcomes achieved through timely diagnosis and consistent monitoring. Tracking the progression of neovascular activity in OCT scans of patients with exudative AMD allows for the development of more personalized and effective treatment plans. This was the focus of the Monitoring Age-related Macular Degeneration Progression in Optical Coherence Tomography (MARIO) challenge, in which we participated. In Task 1, which involved classifying the evolution between two pairs of 2D slices from consecutive OCT acquisitions, we employed a fusion CNN network with model ensembling to further enhance the model's performance. For Task 2, which focused on predicting progression over the next three months based on current exam data, we proposed the Patch Progression Masked Autoencoder that generates an OCT for the next exam and then classifies the evolution between the current OCT and the one generated using our solution from Task 1. The results we achieved allowed us to place in the Top 10 for both tasks. Some team members are part of the same organization as the challenge organizers; therefore, we are not eligible to compete for the prize.
Authors: Zheng Li, Yanming Guo, WenZhe Liu, Xueyi Zhang, Zhaoyun Ding, Long Xu, Mingrui Lao
Abstract: Cross-view geo-localization is a critical task for UAV navigation, event detection, and aerial surveying, as it enables matching between drone-captured and satellite imagery. Most existing approaches embed multi-modal data into a joint feature space to maximize the similarity of paired images. However, these methods typically assume perfect alignment of image pairs during training, which rarely holds true in real-world scenarios. In practice, factors such as urban canyon effects, electromagnetic interference, and adverse weather frequently induce GPS drift, resulting in systematic alignment shifts where only partial correspondences exist between pairs. Despite its prevalence, this source of noisy correspondence has received limited attention in current research. In this paper, we formally introduce and address the Noisy Correspondence on Cross-View Geo-Localization (NC-CVGL) problem, aiming to bridge the gap between idealized benchmarks and practical applications. To this end, we propose PAUL (Partition and Augmentation by Uncertainty Learning), a novel framework that partitions and augments training data based on estimated data uncertainty through uncertainty-aware co-augmentation and evidential co-training. Specifically, PAUL selectively augments regions with high correspondence confidence and utilizes uncertainty estimation to refine feature learning, effectively suppressing noise from misaligned pairs. Distinct from traditional filtering or label correction, PAUL leverages both data uncertainty and loss discrepancy for targeted partitioning and augmentation, thus providing robust supervision for noisy samples. Comprehensive experiments validate the effectiveness of individual components in PAUL,which consistently achieves superior performance over other competitive noisy-correspondence-driven methods in various noise ratios.
Authors: Zhixuan Liang, Yizhuo Li, Tianshuo Yang, Chengyue Wu, Sitong Mao, Liuao Pei, Xiaokang Yang, Jiangmiao Pang, Yao Mu, Ping Luo
Abstract: Vision-Language-Action (VLA) models adapt large vision-language backbones to map images and instructions to robot actions. However, prevailing VLA decoders either generate actions autoregressively in a fixed left-to-right order or attach continuous diffusion or flow matching heads outside the backbone, demanding specialized training and iterative sampling that hinder a unified, scalable architecture. We present Discrete Diffusion VLA, a single-transformer policy that models discretized action chunks with discrete diffusion and is trained with the same cross-entropy objective as the VLM backbone. The design retains diffusion's progressive refinement paradigm while remaining natively compatible with the discrete token interface of VLMs. Our method achieves an adaptive decoding order that resolves easy action elements before harder ones and uses secondary remasking to revisit uncertain predictions across refinement rounds, which improves consistency and enables robust error correction. This unified decoder preserves pretrained vision language priors, supports parallel decoding, breaks the autoregressive bottleneck, and reduces the number of function evaluations. Discrete Diffusion VLA achieves 96.3% avg. SR on LIBERO, 71.2% visual matching on SimplerEnv Fractal and 49.3% overall on SimplerEnv Bridge, improving over both autoregressive and continuous diffusion baselines. These findings indicate that discrete-diffusion action decoder supports precise action modeling and consistent training, laying groundwork for scaling VLA to larger models and datasets.
Authors: Changha Shin, Woong Oh Cho, Seon Joo Kim
Abstract: 360-degree visual content is widely shared on platforms such as YouTube and plays a central role in virtual reality, robotics, and autonomous navigation. However, consumer-grade dual-fisheye systems consistently yield imperfect panoramas due to inherent lens separation and angular distortions. In this work, we introduce a novel calibration framework that incorporates a dual-fisheye camera model into the 3D Gaussian splatting pipeline. Our approach not only simulates the realistic visual artifacts produced by dual-fisheye cameras but also enables the synthesis of seamlessly rendered 360-degree images. By jointly optimizing 3D Gaussian parameters alongside calibration variables that emulate lens gaps and angular distortions, our framework transforms imperfect omnidirectional inputs into flawless novel view synthesis. Extensive evaluations on real-world datasets confirm that our method produces seamless renderings-even from imperfect images-and outperforms existing 360-degree rendering models.
Authors: Yuxin Guo, Teng Wang, Yuying Ge, Shijie Ma, Yixiao Ge, Wei Zou, Ying Shan
Abstract: Recent advances in text-to-audio (TTA) generation excel at synthesizing short audio clips but struggle with long-form narrative audio, which requires temporal coherence and compositional reasoning. To address this gap, we propose AudioStory, a unified framework that integrates large language models (LLMs) with TTA systems to generate structured, long-form audio narratives. AudioStory possesses strong instruction-following reasoning generation capabilities. It employs LLMs to decompose complex narrative queries into temporally ordered sub-tasks with contextual cues, enabling coherent scene transitions and emotional tone consistency. AudioStory has two appealing features: (1) Decoupled bridging mechanism: AudioStory disentangles LLM-diffuser collaboration into two specialized components, i.e., a bridging query for intra-event semantic alignment and a residual query for cross-event coherence preservation. (2) End-to-end training: By unifying instruction comprehension and audio generation within a single end-to-end framework, AudioStory eliminates the need for modular training pipelines while enhancing synergy between components. Furthermore, we establish a benchmark AudioStory-10K, encompassing diverse domains such as animated soundscapes and natural sound narratives. Extensive experiments show the superiority of AudioStory on both single-audio generation and narrative audio generation, surpassing prior TTA baselines in both instruction-following ability and audio fidelity. Our code is available at https://github.com/TencentARC/AudioStory
Authors: Ross J Gardiner, Guillaume Mougeot, Sareh Rowlands, Benno I Simmons, Flemming Helsing, Toke Thomas H{\o}ye
Abstract: Labelling images of Lepidoptera (moths) from automated camera systems is vital for understanding insect declines. However, accurate species identification is challenging due to domain shifts between curated images and noisy field imagery. We propose a lightweight classification approach, combining limited expert-labelled field data with knowledge distillation from the high-performance BioCLIP2 foundation model into a ConvNeXt-tiny architecture. Experiments on 101 Danish moth species from AMI camera systems demonstrate that BioCLIP2 substantially outperforms other methods and that our distilled lightweight model achieves comparable accuracy with significantly reduced computational cost. These insights offer practical guidelines for the development of efficient insect monitoring systems and bridging domain gaps for fine-grained classification.
Authors: Zeyi Sun, Yuhang Cao, Jianze Liang, Qiushi Sun, Ziyu Liu, Zhixiong Zhang, Yuhang Zang, Xiaoyi Dong, Kai Chen, Dahua Lin, Jiaqi Wang
Abstract: Autonomous agents for Graphical User Interfaces (GUIs) face significant challenges in specialized domains such as scientific computing, where both long-horizon planning and precise execution are required. Existing approaches suffer from a trade-off: generalist agents excel at planning but perform poorly in execution, while specialized agents demonstrate the opposite weakness. Recent compositional frameworks attempt to bridge this gap by combining a planner and an actor, but they are typically static and non-trainable, which prevents adaptation from experience. This is a critical limitation given the scarcity of high-quality data in scientific domains. To address these limitations, we introduce CODA, a novel and trainable compositional framework that integrates a generalist planner (Cerebrum) with a specialist executor (Cerebellum), trained via a dedicated two-stage pipeline. In the first stage, Specialization, we apply a decoupled GRPO approach to train an expert planner for each scientific application individually, bootstrapping from a small set of task trajectories. In the second stage, Generalization, we aggregate all successful trajectories from the specialized experts to build a consolidated dataset, which is then used for supervised fine-tuning of the final planner. This equips CODA with both robust execution and cross-domain generalization. Evaluated on four challenging applications from the ScienceBoard benchmark, CODA significantly outperforms baselines and establishes a new state of the art among open-source models.
Authors: Alex Szorkovszky, Rujeena Mathema, Pedro Lencastre, Pedro Lind, Anis Yazidi
Abstract: Although visual search appears largely random, several oculomotor biases exist such that the likelihoods of saccade directions and lengths depend on the previous scan path. Compared to the most recent fixations, the impact of the longer path history is more difficult to quantify. Using the step-selection framework commonly used in movement ecology, and analyzing data from 45-second viewings of "Where's Waldo"?, we report a new memory-dependent effect that also varies significantly between individuals, which we term self-crossing avoidance. This is a tendency for saccades to avoid crossing those earlier in the scan path, and is most evident when both have small amplitudes. We show this by comparing real data to synthetic data generated from a memoryless approximation of the spatial statistics (i.e. a Markovian nonparametric model with a matching distribution of saccade lengths over time). Maximum likelihood fitting indicates that this effect is strongest when including the last $\approx 7$ seconds of a scan path. The effect size is comparable to well-known forms of history dependence such as inhibition of return. A parametric probabilistic model including a self-crossing penalty term was able to reproduce joint statistics of saccade lengths and self-crossings. We also quantified individual strategic differences, and their consistency over the six images viewed per participant, using mixed-effect regressions. Participants with a higher tendency to avoid crossings displayed smaller saccade lengths and shorter fixation durations on average, but did not display more horizontal, vertical, forward or reverse saccades. Together, these results indicate that the avoidance of crossings is a local orienting strategy that facilitates and complements inhibition of return, and hence exploration of visual scenes.
Authors: Gueter Josmy Faure, Min-Hung Chen, Jia-Fong Yeh, Ying Cheng, Hung-Ting Su, Yung-Hao Tang, Shang-Hong Lai, Winston H. Hsu
Abstract: This paper introduces MovieCORE, a novel video question answering (VQA) dataset designed to probe deeper cognitive understanding of movie content. Unlike existing datasets that focus on surface-level comprehension, MovieCORE emphasizes questions that engage System-2 thinking while remaining specific to the video material. We present an innovative agentic brainstorming approach, utilizing multiple large language models (LLMs) as thought agents to generate and refine high-quality question-answer pairs. To evaluate dataset quality, we develop a set of cognitive tests assessing depth, thought-provocation potential, and syntactic complexity. We also propose a comprehensive evaluation scheme for assessing VQA model performance on deeper cognitive tasks. To address the limitations of existing video-language models (VLMs), we introduce an agentic enhancement module, Agentic Choice Enhancement (ACE), which improves model reasoning capabilities post-training by up to 25%. Our work contributes to advancing movie understanding in AI systems and provides valuable insights into the capabilities and limitations of current VQA models when faced with more challenging, nuanced questions about cinematic content. Our project page, dataset and code can be found at https://joslefaure.github.io/assets/html/moviecore.html.
URLs: https://joslefaure.github.io/assets/html/moviecore.html.
Authors: Luvin Munish Ragoo, Ivar Farup, Casper F. Andersen, Graham Finlayson
Abstract: This study investigates the impact of spectral filtering on color-matching functions (CMFs) and its implications for observer variability modeling. We conducted color matching experiments with a single observer, both with and without a spectral filter in front of a bipartite field. Using a novel computational approach, we estimated the filter transmittance and transformation matrix necessary to convert unfiltered CMFs to filtered CMFs. Statistical analysis revealed good agreement between estimated and measured filter characteristics, particularly in central wavelength regions. Applying this methodology to compare between Stiles and Burch 1955 (SB1955) mean observer CMFs and our previously published "ICVIO" mean observer CMFs, we identified a "yellow" (short-wavelength suppressing) filter that effectively transforms between these datasets. This finding aligns with our hypothesis that observed differences between the CMF sets are attributable to age-related lens yellowing (average observer age: 49 years in ICVIO versus 30 years in SB1955). Our approach enables efficient representation of observer variability through a single filter rather than three separate functions, offering potentially reduced experimental overhead while maintaining accuracy in characterizing individual color vision differences.
Authors: Cunmin Zhao, Ziyuan Luo, Guoye Guan, Zelin Li, Yiming Ma, Zhongying Zhao, Renjie Wan
Abstract: 4D live fluorescence microscopy is often compromised by prolonged high intensity illumination which induces photobleaching and phototoxic effects that generate photo-induced artifacts and severely impair image continuity and detail recovery. To address this challenge, we propose the CellINR framework, a case-specific optimization approach based on implicit neural representation. The method employs blind convolution and structure amplification strategies to map 3D spatial coordinates into the high frequency domain, enabling precise modeling and high-accuracy reconstruction of cellular structures while effectively distinguishing true signals from artifacts. Experimental results demonstrate that CellINR significantly outperforms existing techniques in artifact removal and restoration of structural continuity, and for the first time, a paired 4D live cell imaging dataset is provided for evaluating reconstruction performance, thereby offering a solid foundation for subsequent quantitative analyses and biological research. The code and dataset will be public.
Authors: Utsav Ratna Tuladhar, Richard Simon, Doran Mix, Michael Richards
Abstract: Abdominal aortic aneurysms (AAA) pose a significant clinical risk due to their potential for rupture, which is often asymptomatic but can be fatal. Although maximum diameter is commonly used for risk assessment, diameter alone is insufficient as it does not capture the properties of the underlying material of the vessel wall, which play a critical role in determining the risk of rupture. To overcome this limitation, we propose a deep learning-based framework for elasticity imaging of AAAs with 2D ultrasound. Leveraging finite element simulations, we generate a diverse dataset of displacement fields with their corresponding modulus distributions. We train a model with U-Net architecture and normalized mean squared error (NMSE) to infer the spatial modulus distribution from the axial and lateral components of the displacement fields. This model is evaluated across three experimental domains: digital phantom data from 3D COMSOL simulations, physical phantom experiments using biomechanically distinct vessel models, and clinical ultrasound exams from AAA patients. Our simulated results demonstrate that the proposed deep learning model is able to reconstruct modulus distributions, achieving an NMSE score of 0.73\%. Similarly, in phantom data, the predicted modular ratio closely matches the expected values, affirming the model's ability to generalize to phantom data. We compare our approach with an iterative method which shows comparable performance but higher computation time. In contrast, the deep learning method can provide quick and effective estimates of tissue stiffness from ultrasound images, which could help assess the risk of AAA rupture without invasive procedures.
Authors: Pardis Moradbeiki, Nasser Ghadiri, Sayed Jalal Zahabi, Uffe Kock Wiil, Kristoffer Kittelmann Brockhattingen, Ali Ebrahimi
Abstract: Accurate sarcopenia diagnosis via ultrasound remains challenging due to subtle imaging cues, limited labeled data, and the absence of clinical context in most models. We propose MedVQA-TREE, a multimodal framework that integrates a hierarchical image interpretation module, a gated feature-level fusion mechanism, and a novel multi-hop, multi-query retrieval strategy. The vision module includes anatomical classification, region segmentation, and graph-based spatial reasoning to capture coarse, mid-level, and fine-grained structures. A gated fusion mechanism selectively integrates visual features with textual queries, while clinical knowledge is retrieved through a UMLS-guided pipeline accessing PubMed and a sarcopenia-specific external knowledge base. MedVQA-TREE was trained and evaluated on two public MedVQA datasets (VQA-RAD and PathVQA) and a custom sarcopenia ultrasound dataset. The model achieved up to 99% diagnostic accuracy and outperformed previous state-of-the-art methods by over 10%. These results underscore the benefit of combining structured visual understanding with guided knowledge retrieval for effective AI-assisted diagnosis in sarcopenia.
Authors: Xueyang Li, Mingze Jiang, Gelei Xu, Jun Xia, Mengzhao Jia, Danny Chen, Yiyu Shi
Abstract: Agentic AI is advancing rapidly, yet truly autonomous medical-imaging triage, where a system decides when to stop, escalate, or defer under real constraints, remains relatively underexplored. To address this gap, we introduce AT-CXR, an uncertainty-aware agent for chest X-rays. The system estimates per-case confidence and distributional fit, then follows a stepwise policy to issue an automated decision or abstain with a suggested label for human intervention. We evaluate two router designs that share the same inputs and actions: a deterministic rule-based router and an LLM-decided router. Across five-fold evaluation on a balanced subset of NIH ChestX-ray14 dataset, both variants outperform strong zero-shot vision-language models and state-of-the-art supervised classifiers, achieving higher full-coverage accuracy and superior selective-prediction performance, evidenced by a lower area under the risk-coverage curve (AURC) and a lower error rate at high coverage, while operating with lower latency that meets practical clinical constraints. The two routers provide complementary operating points, enabling deployments to prioritize maximal throughput or maximal accuracy. Our code is available at https://github.com/XLIAaron/uncertainty-aware-cxr-agent.
URLs: https://github.com/XLIAaron/uncertainty-aware-cxr-agent.
Authors: Ms. Preeti P. Bhatt, Rakesh R. Savant
Abstract: Steganography is technique of hiding a data under cover media using different steganography tools. Image steganography is hiding of data (Text/Image/Audio/Video) under a cover as Image. This review paper presents classification of image steganography and the comparison of various Image steganography tools using different image formats. Analyzing numerous tools on the basis of Image features and extracting the best one. Some of the tools available in the market were selected based on the frequent use; these tools were tested using the same input on all of them. Specific text was embedded within all host images for each of the six Steganography tools selected. The results of the experiment reveal that all the six tools were relatively performing at the same level, though some software performs better than others through efficiency. And it was based on the image features like size, dimensions, and pixel value and histogram differentiation.
Authors: Marcin Osial, Bartosz W\'ojcik, Bartosz Zieli\'nski, Sebastian Cygert
Abstract: While transfer learning is an advantageous strategy, it overlooks the opportunity to leverage knowledge from numerous available models online. Addressing this multi-source transfer learning problem is a promising path to boost adaptability and cut re-training costs. However, existing approaches are inherently coarse-grained, lacking the necessary precision for granular knowledge extraction and the aggregation efficiency required to fuse knowledge from either a large number of source models or those with high parameter counts. We address these limitations by leveraging Singular Value Decomposition (SVD) to first decompose each source model into its elementary, rank-one components. A subsequent aggregation stage then selects only the most salient components from all sources, thereby overcoming the previous efficiency and precision limitations. To best preserve and leverage the synthesized knowledge base, our method adapts to the target task by fine-tuning only the principal singular values of the merged matrix. In essence, this process only recalibrates the importance of top SVD components. The proposed framework allows for efficient transfer learning, is robust to perturbations both at the input level and in the parameter space (e.g., noisy or pruned sources), and scales well computationally.
Authors: Dikshant Sagar, Kaiwen Yu, Alejandro Yankelevich, Jianming Bian, Pierre Baldi
Abstract: Recent progress in large language models (LLMs) has shown strong potential for multimodal reasoning beyond natural language. In this work, we explore the use of a fine-tuned Vision-Language Model (VLM), based on LLaMA 3.2, for classifying neutrino interactions from pixelated detector images in high-energy physics (HEP) experiments. We benchmark its performance against an established CNN baseline used in experiments like NOvA and DUNE, evaluating metrics such as classification accuracy, precision, recall, and AUC-ROC. Our results show that the VLM not only matches or exceeds CNN performance but also enables richer reasoning and better integration of auxiliary textual or semantic context. These findings suggest that VLMs offer a promising general-purpose backbone for event classification in HEP, paving the way for multimodal approaches in experimental neutrino physics.
Authors: Zhixin Lin, Jungang Li, Shidong Pan, Yibo Shi, Yue Yao, Dongliang Xu
Abstract: Smartphones bring significant convenience to users but also enable devices to extensively record various types of personal information. Existing smartphone agents powered by Multimodal Large Language Models (MLLMs) have achieved remarkable performance in automating different tasks. However, as the cost, these agents are granted substantial access to sensitive users' personal information during this operation. To gain a thorough understanding of the privacy awareness of these agents, we present the first large-scale benchmark encompassing 7,138 scenarios to the best of our knowledge. In addition, for privacy context in scenarios, we annotate its type (e.g., Account Credentials), sensitivity level, and location. We then carefully benchmark seven available mainstream smartphone agents. Our results demonstrate that almost all benchmarked agents show unsatisfying privacy awareness (RA), with performance remaining below 60% even with explicit hints. Overall, closed-source agents show better privacy ability than open-source ones, and Gemini 2.0-flash achieves the best, achieving an RA of 67%. We also find that the agents' privacy detection capability is highly related to scenario sensitivity level, i.e., the scenario with a higher sensitivity level is typically more identifiable. We hope the findings enlighten the research community to rethink the unbalanced utility-privacy tradeoff about smartphone agents. Our code and benchmark are available at https://zhixin-l.github.io/SAPA-Bench.
Authors: Tian Qiu, Alan Zoubi, Yiyuan Lin, Ruiming Du, Lailiang Cheng, Yu Jiang
Abstract: Digital twin applications offered transformative potential by enabling real-time monitoring and robotic simulation through accurate virtual replicas of physical assets. The key to these systems is 3D reconstruction with high geometrical fidelity. However, existing methods struggled under field conditions, especially with sparse and occluded views. This study developed a two-stage framework (DATR) for the reconstruction of apple trees from sparse views. The first stage leverages onboard sensors and foundation models to semi-automatically generate tree masks from complex field images. Tree masks are used to filter out background information in multi-modal data for the single-image-to-3D reconstruction at the second stage. This stage consists of a diffusion model and a large reconstruction model for respective multi view and implicit neural field generation. The training of the diffusion model and LRM was achieved by using realistic synthetic apple trees generated by a Real2Sim data generator. The framework was evaluated on both field and synthetic datasets. The field dataset includes six apple trees with field-measured ground truth, while the synthetic dataset featured structurally diverse trees. Evaluation results showed that our DATR framework outperformed existing 3D reconstruction methods across both datasets and achieved domain-trait estimation comparable to industrial-grade stationary laser scanners while improving the throughput by $\sim$360 times, demonstrating strong potential for scalable agricultural digital twin systems.
Authors: Hail Song, Seokhwan Yang, Woontack Woo
Abstract: We present a fast and efficient method for transferring facial textures onto SMPL-X-based full-body avatars. Unlike conventional affine-transform methods that are slow and prone to visual artifacts, our method utilizes a barycentric UV conversion technique. Our approach precomputes the entire UV mapping into a single transformation matrix, enabling texture transfer in a single operation. This results in a speedup of over 7000x compared to the baseline, while also significantly improving the final texture quality by eliminating boundary artifacts. Through quantitative and qualitative evaluations, we demonstrate that our method offers a practical solution for personalization in immersive XR applications. The code is available online.
Authors: Subhrojyoti Mukherjee, Manoranjan Mohanty
Abstract: Fake images in selfie banking are increasingly becoming a threat. Previously, it was just Photoshop, but now deep learning technologies enable us to create highly realistic fake identities, which fraudsters exploit to bypass biometric systems such as facial recognition in online banking. This paper explores the use of an already established forensic recognition system, previously used for picture camera localization, in deepfake detection.
Authors: Sena Ishii, Akash Chikhalikar, Ankit A. Ravankar, Jose Victorio Salazar Luces, Yasuhisa Hirata
Abstract: We present a novel framework for estimating accident-prone regions in everyday indoor scenes, aimed at improving real-time risk awareness in service robots operating in human-centric environments. As robots become integrated into daily life, particularly in homes, the ability to anticipate and respond to environmental hazards is crucial for ensuring user safety, trust, and effective human-robot interaction. Our approach models object-level risk and context through a semantic graph-based propagation algorithm. Each object is represented as a node with an associated risk score, and risk propagates asymmetrically from high-risk to low-risk objects based on spatial proximity and accident relationship. This enables the robot to infer potential hazards even when they are not explicitly visible or labeled. Designed for interpretability and lightweight onboard deployment, our method is validated on a dataset with human-annotated risk regions, achieving a binary risk detection accuracy of 75%. The system demonstrates strong alignment with human perception, particularly in scenes involving sharp or unstable objects. These results underline the potential of context-aware risk reasoning to enhance robotic scene understanding and proactive safety behaviors in shared human-robot spaces. This framework could serve as a foundation for future systems that make context-driven safety decisions, provide real-time alerts, or autonomously assist users in avoiding or mitigating hazards within home environments.
Authors: Mohammed Rakibul Hasan, Rafi Majid, Ahanaf Tahmid
Abstract: In this paper, we introduce Bangla-Bayanno, an open-ended Visual Question Answering (VQA) Dataset in Bangla, a widely used, low-resource language in multimodal AI research. The majority of existing datasets are either manually annotated with an emphasis on a specific domain, query type, or answer type or are constrained by niche answer formats. In order to mitigate human-induced errors and guarantee lucidity, we implemented a multilingual LLM-assisted translation refinement pipeline. This dataset overcomes the issues of low-quality translations from multilingual sources. The dataset comprises 52,650 question-answer pairs across 4750+ images. Questions are classified into three distinct answer types: nominal (short descriptive), quantitative (numeric), and polar (yes/no). Bangla-Bayanno provides the most comprehensive open-source, high-quality VQA benchmark in Bangla, aiming to advance research in low-resource multimodal learning and facilitate the development of more inclusive AI systems.
Authors: Davorin Mili\v{c}evi\'c, Ratko Grbi\'c
Abstract: Deep Convolutional Neural Networks (CNNs) achieve high accuracy but often rely on purely global, gradient-based optimisation, which can lead to overfitting, redundant filters, and reduced interpretability. To address these limitations, we propose NM-Hebb, a two-phase training framework that integrates neuro-inspired local plasticity with distance-aware supervision. Phase 1 extends standard supervised training by jointly optimising a cross-entropy objective with two biologically inspired mechanisms: (i) a Hebbian regulariser that aligns the spatial mean of activations with the mean of the corresponding convolutional filter weights, encouraging structured, reusable primitives; and (ii) a learnable neuromodulator that gates an elastic-weight-style consolidation loss, preserving beneficial parameters without freezing the network. Phase 2 fine-tunes the backbone with a pairwise metric-learning loss, explicitly compressing intra-class distances and enlarging inter-class margins in the embedding space. Evaluated on CIFAR-10, CIFAR-100, and TinyImageNet across five backbones (ResNet-18, VGG-11, MobileNet-v2, EfficientNet-V2, DenseNet-121), NM-Hebb achieves consistent gains over baseline and other methods: Top-1 accuracy improves by +2.0-10.0 pp (CIFAR-10), +2.0-9.0 pp (CIFAR-100), and up to +4.3-8.9 pp (TinyImageNet), with Normalised Mutual Information (NMI) increased by up to +0.15. Qualitative visualisations and filter-level analyses further confirm that NM-Hebb produces more structured and selective features, yielding tighter and more interpretable class clusters. Overall, coupling local Hebbian plasticity with metric-based fine-tuning yields CNNs that are not only more accurate but also more interpretable, offering practical benefits for resource-constrained and safety-critical AI deployments.
Authors: Quanfeng Lu, Zhantao Ma, Shuai Zhong, Jin Wang, Dahai Yu, Michael K. Ng, Ping Luo
Abstract: The rapid advancement of large vision language models (LVLMs) and agent systems has heightened interest in mobile GUI agents that can reliably translate natural language into interface operations. Existing single-agent approaches, however, remain limited by structural constraints. Although multi-agent systems naturally decouple different competencies, recent progress in multi-agent reinforcement learning (MARL) has often been hindered by inefficiency and remains incompatible with current LVLM architectures. To address these challenges, we introduce SWIRL, a staged workflow for interleaved reinforcement learning designed for multi-agent systems. SWIRL reformulates MARL into a sequence of single-agent reinforcement learning tasks, updating one agent at a time while keeping the others fixed. This formulation enables stable training and promotes efficient coordination across agents. Theoretically, we provide a stepwise safety bound, a cross-round monotonic improvement theorem, and convergence guarantees on return, ensuring robust and principled optimization. In application to mobile GUI control, SWIRL instantiates a Navigator that converts language and screen context into structured plans, and an Interactor that grounds these plans into executable atomic actions. Extensive experiments demonstrate superior performance on both high-level and low-level GUI benchmarks. Beyond GUI tasks, SWIRL also demonstrates strong capability in multi-agent mathematical reasoning, underscoring its potential as a general framework for developing efficient and robust multi-agent systems.
Authors: Chengzu Li, Wenshan Wu, Huanyu Zhang, Qingtao Li, Zeyu Gao, Yan Xia, Jos\'e Hern\'andez-Orallo, Ivan Vuli\'c, Furu Wei
Abstract: For human cognitive process, spatial reasoning and perception are closely entangled, yet the nature of this interplay remains underexplored in the evaluation of multimodal large language models (MLLMs). While recent MLLM advancements show impressive performance on reasoning, their capacity for human-like spatial cognition remains an open question. In this work, we introduce a systematic evaluation framework to assess the spatial reasoning abilities of state-of-the-art MLLMs relative to human performance. Central to our work is 11Plus-Bench, a high-quality benchmark derived from realistic standardized spatial aptitude tests. 11Plus-Bench also features fine-grained expert annotations of both perceptual complexity and reasoning process, enabling detailed instance-level analysis of model behavior. Through extensive experiments across 14 MLLMs and human evaluation, we find that current MLLMs exhibit early signs of spatial cognition. Despite a large performance gap compared to humans, MLLMs' cognitive profiles resemble those of humans in that cognitive effort correlates strongly with reasoning-related complexity. However, instance-level performance in MLLMs remains largely random, whereas human correctness is highly predictable and shaped by abstract pattern complexity. These findings highlight both emerging capabilities and limitations in current MLLMs' spatial reasoning capabilities and provide actionable insights for advancing model design.
Authors: Maud Biquard, Marie Chabert, Florence Genin, Christophe Latry, Thomas Oberlin
Abstract: Regularization of inverse problems is of paramount importance in computational imaging. The ability of neural networks to learn efficient image representations has been recently exploited to design powerful data-driven regularizers. While state-of-the-art plug-and-play (PnP) methods rely on an implicit regularization provided by neural denoisers, alternative Bayesian approaches consider Maximum A Posteriori (MAP) estimation in the latent space of a generative model, thus with an explicit regularization. However, state-of-the-art deep generative models require a huge amount of training data compared to denoisers. Besides, their complexity hampers the optimization involved in latent MAP derivation. In this work, we first propose to use compressive autoencoders instead. These networks, which can be seen as variational autoencoders with a flexible latent prior, are smaller and easier to train than state-of-the-art generative models. As a second contribution, we introduce the Variational Bayes Latent Estimation (VBLE) algorithm, which performs latent estimation within the framework of variational inference. Thanks to a simple yet efficient parameterization of the variational posterior, VBLE allows for fast and easy (approximate) posterior sampling. Experimental results on image datasets BSD and FFHQ demonstrate that VBLE reaches similar performance as state-of-the-art PnP methods, while being able to quantify uncertainties significantly faster than other existing posterior sampling techniques. The code associated to this paper is available in https://github.com/MaudBqrd/VBLE.
Authors: Nikita Gabdullin
Abstract: Autoencoders (AE) are simple yet powerful class of neural networks that compress data by projecting input into low-dimensional latent space (LS). Whereas LS is formed according to the loss function minimization during training, its properties and topology are not controlled directly. In this paper we focus on AE LS properties and propose two methods for obtaining LS with desired topology, called LS configuration. The proposed methods include loss configuration using a geometric loss term that acts directly in LS, and encoder configuration. We show that the former allows to reliably obtain LS with desired configuration by defining the positions and shapes of LS clusters for supervised AE (SAE). Knowing LS configuration allows to define similarity measure in LS to predict labels or estimate similarity for multiple inputs without using decoders or classifiers. We also show that this leads to more stable and interpretable training. We show that SAE trained for clothes texture classification using the proposed method generalizes well to unseen data from LIP, Market1501, and WildTrack datasets without fine-tuning, and even allows to evaluate similarity for unseen classes. We further illustrate the advantages of pre-configured LS similarity estimation with cross-dataset searches and text-based search using a text query without language models.
Authors: Haonan Han, Rui Yang, Huan Liao, Jiankai Xing, Zunnan Xu, Xiaoming Yu, Junwei Zha, Xiu Li, Wanhua Li
Abstract: Traditional image-to-3D models often struggle with scenes containing multiple objects due to biases and occlusion complexities. To address this challenge, we present REPARO, a novel approach for compositional 3D asset generation from single images. REPARO employs a two-step process: first, it extracts individual objects from the scene and reconstructs their 3D meshes using off-the-shelf image-to-3D models; then, it optimizes the layout of these meshes through differentiable rendering techniques, ensuring coherent scene composition. By integrating optimal transport-based long-range appearance loss term and high-level semantic loss term in the differentiable rendering, REPARO can effectively recover the layout of 3D assets. The proposed method can significantly enhance object independence, detail accuracy, and overall scene coherence. Extensive evaluation of multi-object scenes demonstrates that our REPARO offers a comprehensive approach to address the complexities of multi-object 3D scene generation from single images.
Authors: Mingyuan Wu, Zichuan Liu, Haozhen Zheng, Hongpeng Guo, Bo Chen, Xin Lu, Klara Nahrstedt
Abstract: Efficient single instance segmentation is essential for unlocking features in the mobile imaging applications, such as capture or editing. Existing on-the-fly mobile imaging applications scope the segmentation task to portraits or the salient subject due to the computational constraints. Instance segmentation, despite its recent developments towards efficient networks, is still heavy due to the cost of computation on the entire image to identify all instances. To address this, we propose and formulate a one tap driven single instance segmentation task that segments a single instance selected by a user via a positive tap. This task, in contrast to the broader task of segmenting anything as suggested in the Segment Anything Model \cite{sam}, focuses on efficient segmentation of a single instance specified by the user. To solve this problem, we present TraceNet, which explicitly locates the selected instance by way of receptive field tracing. TraceNet identifies image regions that are related to the user tap and heavy computations are only performed on selected regions of the image. Therefore overall computation cost and memory consumption are reduced during inference. We evaluate the performance of TraceNet on instance IoU average over taps and the proportion of the region that a user tap can fall into for a high-quality single-instance mask. Experimental results on MS-COCO and LVIS demonstrate the effectiveness and efficiency of the proposed approach. TraceNet can jointly achieve the efficiency and interactivity, filling in the gap between needs for efficient mobile inference and recent research trend towards multimodal and interactive segmentation models.
Authors: Pedro R. A. S. Bassi, Haydr A. H. Ali, Andrea Cavalli, Sergio Decherchi
Abstract: Application of Artificial Intelligence (AI) in critical domains, like the medical one, is often hampered by shortcut learning, which hinders AI generalization to diverse hospitals and patients. Shortcut learning can be caused, for example, by background biases -- features in image backgrounds that are spuriously correlated to classification labels (e.g., words in X-rays). To mitigate the influence of image background and foreground bias on AI, we introduce a new training paradigm, dubbed Training with Explanations Alone (TEA). TEA trains a classifier (TEA student) only by making its explanation heatmaps match target heatmaps from a larger teacher model. By learning from its explanation heatmaps, the TEA student pays attention to the same image features as the teacher. For example, a teacher uses a large segmenter to remove image backgrounds before classification, thus ignoring background bias. By learning from the teacher's explanation heatmaps, the TEA student learns to also ignore backgrounds -- but it does not need a segmenter. With different teachers, the TEA student can also resist bias in the image foreground. Surprisingly, by training with heatmaps alone the student output naturally matches the teacher output -- with no loss function applied to the student output. We compared the TEA student against 14 state-of-the-art methods in 5 datasets with strong background or foreground bias, including Waterbirds and an X-Ray dataset for COVID-19 and pneumonia classification. The TEA student had better resistance to bias, strongly surpassing state-of-the-art methods, and generalizing better to hospitals not seen in training.
Authors: Jingwei Guo, Yitai Cheng, Meihui Wang, Ilya Ilyankou, Natchapon Jongwiriyanurak, Xiaowei Gao, Nicola Christie, James Haworth
Abstract: Cyclists face a disproportionate risk of injury, yet conventional crash records are too limited to reconstruct the circumstances of incidents or to diagnose risk at the finer spatial and temporal detail needed for targeted interventions. Recently, naturalistic studies have gained traction as a way to capture the complex behavioural and infrastructural factors that contribute to crashes. These approaches typically involve the collection and analysis of video data. A video promising format is panoramic video, which can record 360-degree views around a rider. However, its use is limited by severe distortions, large numbers of small objects and boundary continuity. This study addresses these challenges by proposing a novel three-step framework: (1) enhancing object detection accuracy on panoramic imagery by segmenting and projecting the original 360-degree images into four perspective sub-images, thus reducing distortion; (2) modifying multi-object tracking models to incorporate boundary continuity and object category information for improved tracking consistency; and (3) validating the proposed approach through a real-world application focused on detecting overtaking manoeuvres by vehicles around cyclists. The methodology is evaluated using panoramic videos recorded by cyclists on London's roadways under diverse conditions. Experimental results demonstrate notable improvements over baseline methods, achieving higher average precision across varying image resolutions. Moreover, the enhanced tracking approach yields a 3.0% increase in multi-object tracking accuracy and a 4.6% improvement in identification F-score. The overtaking detection task achieves a high F-score of 0.81, illustrating the practical effectiveness of the proposed method in real-world cycling safety scenarios. The code is available on GitHub (https://github.com/SpaceTimeLab/360_object_tracking) to ensure reproducibility.
Authors: Ruixiang Jiang, Changwen Chen
Abstract: Artistic styles are defined by both their structural and appearance elements. Existing neural stylization techniques primarily focus on transferring appearance-level features such as color and texture, often neglecting the equally crucial aspect of structural stylization. To address this gap, we introduce \textbf{DiffArtist}, the first 2D stylization method to offer fine-grained, simultaneous control over both structure and appearance style strength. This dual controllability is achieved by representing structure and appearance generation as separate diffusion processes, necessitating no further tuning or additional adapters. To properly evaluate this new capability of dual stylization, we further propose a Multimodal LLM-based stylization evaluator that aligns significantly better with human preferences than existing metrics. Extensive analysis shows that DiffArtist achieves superior style fidelity and dual-controllability compared to state-of-the-art methods. Its text-driven, training-free design and unprecedented dual controllability make it a powerful and interactive tool for various creative applications. Project homepage: https://diffusionartist.github.io.
Authors: Jingyun Wang, Guoliang Kang
Abstract: Recent works utilize CLIP to perform the challenging unsupervised semantic segmentation task where only images without annotations are available. However, we observe that when adopting CLIP to such a pixel-level understanding task, unexpected bias (including class-preference bias and space-preference bias) occurs. Previous works don't explicitly model the bias, which largely constrains the segmentation performance. In this paper, we propose to explicitly model and rectify the bias existing in CLIP to facilitate the unsupervised semantic segmentation task. Specifically, we design a learnable "Reference" prompt to encode class-preference bias and a projection of the positional embedding in the vision transformer to encode space-preference bias respectively. To avoid interference, two kinds of biases are firstly independently encoded into different features, i.e., the Reference feature and the positional feature. Via a matrix multiplication between the Reference feature and the positional feature, a bias logit map is generated to explicitly represent two kinds of biases. Then we rectify the logits of CLIP via a simple element-wise subtraction. To make the rectified results smoother and more contextual, we design a mask decoder which takes the feature of CLIP and the rectified logits as input and outputs a rectified segmentation mask with the help of Gumbel-Softmax operation. A contrastive loss based on the masked visual features and the text features of different classes is imposed, which makes the bias modeling and rectification process meaningful and effective. Extensive experiments on various benchmarks including PASCAL VOC, PASCAL Context, ADE20K, Cityscapes, and COCO Stuff demonstrate that our method performs favorably against previous state-of-the-arts. The implementation is available at: https://github.com/dogehhh/ReCLIP.
Authors: Haozhan Shen, Kangjia Zhao, Tiancheng Zhao, Ruochen Xu, Zilun Zhang, Mingwei Zhu, Jianwei Yin
Abstract: Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in vision-language understanding. Recently, with the integration of test-time scaling techniques, these models have also shown strong potential in visual reasoning. However, most existing reasoning approaches remain text-level in nature: MLLMs are prompted to explore various combinations of textual tokens via their underlying language model, while the visual input remains fixed throughout the reasoning process. This paradigm limits the model's ability to fully exploit rich visual information, particularly when dealing with images containing numerous fine-grained elements. In such cases, vision-level reasoning becomes crucial - where models dynamically zoom into specific regions of the image to gather detailed visual cues necessary for accurate decision-making. In this paper, we propose Zoom Eye, a training-free, model-agnostic tree search algorithm tailored for vision-level reasoning. Zoom Eye treats an image as a hierarchical tree structure, where each child node represents a zoomed-in sub-region of its parent, and the root corresponds to the full image. The algorithm enables MLLMs to simulate human-like zooming behavior by navigating from root to leaf nodes in search of task-relevant visual evidence. We experiment on a series of high-resolution benchmarks and the results demonstrate that Zoom Eye consistently improves the performance of multiple MLLMs by a large margin (e.g., InternVL2.5-8B increases by 15.71% and 17.69% on HR-Bench) and also enables small 3-8B MLLMs to outperform strong large models such as GPT-4o. Code: https://github.com/om-ai-lab/ZoomEye
Authors: Zeyi Sun, Ziyang Chu, Pan Zhang, Tong Wu, Xiaoyi Dong, Yuhang Zang, Yuanjun Xiong, Dahua Lin, Jiaqi Wang
Abstract: In-context generation is a key component of large language models' (LLMs) open-task generalization capability. By leveraging a few examples as context, LLMs can perform both in-domain and out-of-domain tasks. Recent advancements in auto-regressive vision-language models (VLMs) built upon LLMs have showcased impressive performance in text-to-image generation. However, the potential of in-context learning for general image generation tasks remains largely unexplored. To address this, we introduce X-Prompt, a purely auto-regressive large-vision language model designed to deliver competitive performance across a wide range of both seen and unseen image generation tasks, all within a unified in-context learning framework. X-Prompt incorporates a specialized design that efficiently compresses valuable features from in-context examples, supporting longer in-context token sequences and improving its ability to generalize to unseen tasks. A unified training task for both text and image prediction enables X-Prompt to handle general image generation with enhanced task awareness from in-context examples. Extensive experiments validate the model's performance across diverse seen image generation tasks and its capacity to generalize to previously unseen tasks.
Authors: Kuntao Xiao, Xiongfei Wang, Pengfei Teng, Yi Sun, Yong Zhang, Wanli Yang, Zikang Xu, Liang Zhang, Hanyang Dong, Guoming Luan, Shurong Sheng
Abstract: The analysis of interictal epileptiform discharges (IEDs) in magnetoencephalography (MEG) or electroencephalogram (EEG) recordings represents a critical component in the diagnosis of epilepsy. However, manual analysis of these IEDs, which appear as epileptic spikes, from the large amount of MEG/EEG data is labor intensive and requires high expertise. Although automated methods have been developed to address this challenge, current approaches fail to fully emulate clinical experts' diagnostic intelligence in two key aspects: (1) their analysis on the input signals is limited to short temporal windows matching individual spike durations, missing the extended contextual patterns clinicians use to assess significance; and (2) they fail to adequately capture the dipole patterns with simultaneous positive-negative potential distributions across adjacent sensors that serve as clinicians' key diagnostic criterion for IED identification. To bridge this artificial-human intelligence gap, we propose a novel deep learning framework LV-CadeNet that integrates two key innovations: (1) a Long-View morphological feature representation that mimics expert clinicians' comprehensive assessment of both local spike characteristics and long-view contextual information, and (2) a hierarchical Encoder-Decoder NETwork that employs Convolution-Attention blocks for multi-scale spatiotemporal feature learning with progressive abstraction. Extensive evaluations confirm the superior performance of LV-CadeNet, which outperforms six state-of-the-art methods in EEG spike classification on TUEV, the largest public EEG spike dataset. Additionally, LV-CadeNet attains a significant improvement of 13.58% in balanced accuracy over the leading baseline for MEG spike detection on a clinical MEG dataset from Sanbo Brain Hospital, Capital Medical University.
Authors: Peirong Zhang, Lianwen Jin
Abstract: Currently, the prevalence of online handwriting has spurred a critical need for effective retrieval systems to accurately search relevant handwriting instances from specific writers, known as online writer retrieval. Despite the growing demand, this field suffers from a scarcity of well-established methodologies and public large-scale datasets. This paper tackles these challenges with a focus on Chinese handwritten phrases. First, we propose DOLPHIN, a novel retrieval model designed to enhance handwriting representations through synergistic temporal-frequency analysis. For frequency feature learning, we propose the HFGA block, which performs gated cross-attention between the vanilla temporal handwriting sequence and its high-frequency sub-bands to amplify salient writing details. For temporal feature learning, we propose the CAIR block, tailored to promote channel interaction and reduce channel redundancy. Second, to address data deficit, we introduce OLIWER, a large-scale online writer retrieval dataset encompassing over 670,000 Chinese handwritten phrases from 1,731 individuals. Through extensive evaluations, we demonstrate the superior performance of DOLPHIN over existing methods. In addition, we explore cross-domain writer retrieval and reveal the pivotal role of increasing feature alignment in bridging the distributional gap between different handwriting data. Our findings emphasize the significance of point sampling frequency and pressure features in improving handwriting representation quality and retrieval performance. Code and dataset are available at https://github.com/SCUT-DLVCLab/DOLPHIN.
Authors: Xianfeng Song, Yi Zou, Zheng Shi, Zheng Liu
Abstract: Feature-based image matching has extensive applications in computer vision. Keypoints detected in images can be naturally represented as graph structures, and Graph Neural Networks (GNNs) have been shown to outperform traditional deep learning techniques. Consequently, the paradigm of image matching via GNNs has gained significant prominence in recent academic research. In this paper, we first introduce an innovative adaptive graph construction method that utilizes a filtering mechanism based on distance and dynamic threshold similarity. This method dynamically adjusts the criteria for incorporating new vertices based on the characteristics of existing vertices, allowing for the construction of more precise and robust graph structures while avoiding redundancy. We further combine the vertex processing capabilities of GNNs with the global awareness capabilities of Transformers to enhance the model's representation of spatial and feature information within graph structures. This hybrid model provides a deeper understanding of the interrelationships between vertices and their contributions to the matching process. Additionally, we employ the Sinkhorn algorithm to iteratively solve for optimal matching results. Finally, we validate our system using extensive image datasets and conduct comprehensive comparative experiments. Experimental results demonstrate that our system achieves an average improvement of 3.8x-40.3x in overall matching performance. Additionally, the number of vertices and edges significantly impacts training efficiency and memory usage; therefore, we employ multi-GPU technology to accelerate the training process. Our code is available at https://github.com/songxf1024/GIMS.
Authors: Felix Duelmer, Mohammad Farid Azampour, Magdalena Wysocki, Nassir Navab
Abstract: Traditional ultrasound simulators solve the wave equation to model pressure distribution fields, achieving high accuracy but requiring significant computational time and resources. To address this, ray tracing approaches have been introduced, modeling wave propagation as rays interacting with boundaries and scatterers. However, existing models simplify ray propagation, generating echoes at interaction points without considering return paths to the sensor. This can result in unrealistic artifacts and necessitates careful scene tuning for plausible results. We propose a novel ultrasound simulation pipeline that utilizes a ray tracing algorithm to generate echo data, tracing each ray from the transducer through the scene and back to the sensor. To replicate advanced ultrasound imaging, we introduce a ray emission scheme optimized for plane wave imaging, incorporating delay and steering capabilities. Furthermore, we integrate a standard signal processing pipeline to simulate end-to-end ultrasound image formation. We showcase the efficacy of the proposed pipeline by modeling synthetic scenes featuring highly reflective objects, such as bones. In doing so, our proposed approach, UltraRay, not only enhances the overall visual quality but also improves the realism of the simulated images by accurately capturing secondary reflections and reducing unnatural artifacts. By building on top of a differentiable framework, the proposed pipeline lays the groundwork for a fast and differentiable ultrasound simulation tool necessary for gradient-based optimization, enabling advanced ultrasound beamforming strategies, neural network integration, and accurate inverse scene reconstruction.
Authors: Junsung Park, Jungbeom Lee, Jongyoon Song, Sangwon Yu, Dahuin Jung, Sungroh Yoon
Abstract: While CLIP has significantly advanced multimodal understanding by bridging vision and language, the inability to grasp negation - such as failing to differentiate concepts like "parking" from "no parking" - poses substantial challenges. By analyzing the data used in the public CLIP model's pre-training, we posit this limitation stems from a lack of negation-inclusive data. To address this, we introduce data generation pipelines that employ a large language model (LLM) and a multimodal LLM to produce negation-inclusive captions. Fine-tuning CLIP with data generated from our pipelines, we develop NegationCLIP, which enhances negation awareness while preserving the generality. Moreover, to enable a comprehensive evaluation of negation understanding, we propose NegRefCOCOg-a benchmark tailored to test VLMs' ability to interpret negation across diverse expressions and positions within a sentence. Experiments on various CLIP architectures validate the effectiveness of our data generation pipelines in enhancing CLIP's ability to perceive negation accurately. Additionally, NegationCLIP's enhanced negation awareness has practical applications across various multimodal tasks, demonstrated by performance gains in text-to-image generation and referring image segmentation.
Authors: Matt C. Bendel, Saurav K. Shastri, Rizwan Ahmad, Philip Schniter
Abstract: Imaging inverse problems can be solved in an unsupervised manner using pre-trained diffusion models, but doing so requires approximating the gradient of the measurement-conditional score function in the diffusion reverse process. We show that the approximations produced by existing methods are relatively poor, especially early in the reverse process, and so we propose a new approach that iteratively reestimates and "renoises" the estimate several times per diffusion step. This iterative approach, which we call Fast Iterative REnoising (FIRE), injects colored noise that is shaped to ensure that the pre-trained diffusion model always sees white noise, in accordance with how it was trained. We then embed FIRE into the DDIM reverse process and show that the resulting "DDfire" offers state-of-the-art accuracy and runtime on several linear inverse problems, as well as phase retrieval. Our implementation is at https://github.com/matt-bendel/DDfire
Authors: Mohamed Youssef, Jian Peng, Oliver Bimber
Abstract: Access to below-canopy volumetric vegetation data is crucial for understanding ecosystem dynamics. We address the long-standing limitation of remote sensing to penetrate deep into dense canopy layers. LiDAR and radar are currently considered the primary options for measuring 3D vegetation structures, while cameras can only extract the reflectance and depth of top layers. Using conventional, high-resolution aerial images, our approach allows sensing deep into self-occluding vegetation volumes, such as forests. It is similar in spirit to the imaging process of wide-field microscopy, but can handle much larger scales and strong occlusion. We scan focal stacks by synthetic-aperture imaging with drones and reduce out-of-focus signal contributions using pre-trained 3D convolutional neural networks with mean squared error (MSE) as the loss function. The resulting volumetric reflectance stacks contain low-frequency representations of the vegetation volume. Combining multiple reflectance stacks from various spectral channels provides insights into plant health, growth, and environmental conditions throughout the entire vegetation volume. Compared with simulated ground truth, our correction leads to ~x7 average improvements (min: ~x2, max: ~x12) for forest densities of 220 trees/ha - 1680 trees/ha. In our field experiment, we achieved an MSE of 0.05 when comparing with the top-vegetation layer that was measured with classical multispectral aerial imaging.
Authors: Hongseok Oh, Wonseok Hwang
Abstract: Recently, Large Vision-Language Models (LVLMs) show remarkable performance across various domains. However, these models suffer from object hallucination. This study revisits the previous claim that the cause of such hallucinations lies in the limited representational capacity of the vision encoder. Our analysis implies that the capacity of the vision encoder is not necessarily a major limiting factor in detecting object hallucination. Based on this insight, we propose Fine-grained CLIPScore (F-CLIPScore), a simple yet effective evaluation metric that enhances object-level granularity by incorporating text embeddings at the noun level. Evaluations on the OHD-Caps benchmark show that F-CLIPScore significantly outperforms conventional CLIPScore in accuracy by a large margin of \textbf{39.6\%} without additional training. We further demonstrate that F-CLIPScore-based data filtering reduces object hallucination in LVLM (4.9\% in POPE).
Authors: Patryk Rygiel, Julian Suk, Kak Khee Yeung, Christoph Brune, Jelmer M. Wolterink
Abstract: Hemodynamic parameters such as pressure and wall shear stress play an important role in diagnosis, prognosis, and treatment planning in cardiovascular diseases. These parameters can be accurately computed using computational fluid dynamics (CFD), but CFD is computationally intensive. Hence, deep learning methods have been adopted as a surrogate to rapidly estimate CFD outcomes. A drawback of such data-driven models is the need for time-consuming reference CFD simulations for training. In this work, we introduce an active learning framework to reduce the number of CFD simulations required for the training of surrogate models, lowering the barriers to their deployment in new applications. We propose three distinct querying strategies to determine for which unlabeled samples CFD simulations should be obtained. These querying strategies are based on geometrical variance, ensemble uncertainty, and adherence to the physics governing fluid dynamics. We benchmark these methods on velocity field estimation in synthetic coronary artery bifurcations and find that they allow for substantial reductions in annotation cost. Notably, we find that our strategies reduce the number of samples required by up to 50% and make the trained models more robust to difficult cases. Our results show that active learning is a feasible strategy to increase the potential of deep learning-based CFD surrogates.
Authors: Tieqiao Wang, Sinisa Todorovic
Abstract: Most recent work on action segmentation relies on pre-computed frame features from models trained on other tasks and typically focuses on framewise encoding and labeling without explicitly modeling action segments. To overcome these limitations, we introduce the End-to-End Action Segmentation Transformer (EAST), which processes raw video frames directly -- eliminating the need for pre-extracted features and enabling true end-to-end training. Our contributions are as follows: (1) a lightweight adapter design for effective fine-tuning of large backbones; (2) an efficient segmentation-by-detection framework for leveraging action proposals predicted over a coarsely downsampled video; and (3) a novel action-proposal-based data augmentation strategy. EAST achieves SOTA performance on standard benchmarks, including GTEA, 50Salads, Breakfast, and Assembly-101.
Authors: Hao Cheng, Erjia Xiao, Yichi Wang, Lingfeng Zhang, Qiang Zhang, Jiahang Cao, Kaidi Xu, Mengshu Sun, Xiaoshuai Hao, Jindong Gu, Renjing Xu
Abstract: Current Cross-Modality Generation Models (GMs) demonstrate remarkable capabilities in various generative tasks. Given the ubiquity and information richness of vision modality inputs in real-world scenarios, Cross-Vision tasks, encompassing Vision-Language Perception (VLP) and Image-to-Image (I2I), have attracted significant attention. Large Vision Language Models (LVLMs) and I2I Generation Models (GMs) are employed to handle VLP and I2I tasks, respectively. Previous research indicates that printing typographic words into input images significantly induces LVLMs and I2I GMs to produce disruptive outputs that are semantically aligned with those words. Additionally, visual prompts, as a more sophisticated form of typography, are also revealed to pose security risks to various applications of cross-vision tasks. However, the specific characteristics of the threats posed by visual prompts remain underexplored. In this paper, to comprehensively investigate the performance impact induced by Typographic Visual Prompt Injection (TVPI) in various LVLMs and I2I GMs, we propose the Typographic Visual Prompts Injection Dataset and thoroughly evaluate the TVPI security risks on various open-source and closed-source LVLMs and I2I GMs under visual prompts with different target semantics, deepening the understanding of TVPI threats.
Authors: Jaywon Koo, Jefferson Hernandez, Moayed Haji-Ali, Ziyan Yang, Vicente Ordonez
Abstract: Evaluating text-to-image and text-to-video models is challenging due to a fundamental disconnect: established metrics fail to jointly measure visual quality and semantic alignment with text, leading to a poor correlation with human judgments. To address this critical issue, we propose cFreD, a general metric based on a Conditional Fr\'echet Distance that unifies the assessment of visual fidelity and text-prompt consistency into a single score. Existing metrics such as Fr\'echet Inception Distance (FID) capture image quality but ignore text conditioning while alignment scores such as CLIPScore are insensitive to visual quality. Furthermore, learned preference models require constant retraining and are unlikely to generalize to novel architectures or out-of-distribution prompts. Through extensive experiments across multiple recently proposed text-to-image models and diverse prompt datasets, cFreD exhibits a higher correlation with human judgments compared to statistical metrics , including metrics trained with human preferences. Our findings validate cFreD as a robust, future-proof metric for the systematic evaluation of text conditioned models, standardizing benchmarking in this rapidly evolving field. We release our evaluation toolkit and benchmark.
Authors: Shuhao Kang, Martin Y. Liao, Yan Xia, Olaf Wysocki, Boris Jutzi, Daniel Cremers
Abstract: LiDAR place recognition is a critical capability for autonomous navigation and cross-modal localization in large-scale outdoor environments. Existing approaches predominantly depend on pre-built 3D dense maps or aerial imagery, which impose significant storage overhead and lack real-time adaptability. In this paper, we propose OPAL, a novel framework for LiDAR place recognition that leverages OpenStreetMap (OSM) as a lightweight and up-to-date prior. Our key innovation lies in bridging the domain disparity between sparse LiDAR scans and structured OSM data through two carefully designed components. First, a cross-modal visibility mask that identifies observable regions from both modalities to guide feature alignment. Second, an adaptive radial fusion module that dynamically consolidates radial features into discriminative global descriptors. Extensive experiments on KITTI and KITTI-360 datasets demonstrate OPAL's superiority, achieving 15.98% higher recall at 1m threshold for top-1 retrieved matches, along with 12x faster inference speed compared to the state-of-the-art approach. Code and data are publicly available at: https://github.com/kang-1-2-3/OPAL.
Authors: Pengfei Zhang, Shouqing Jia
Abstract: Denoising Diffusion Probabilistic Models (DDPM) process images as a whole. Since adjacent pixels are highly likely to belong to the same object, we propose the Heat Diffusion Model (HDM) to further preserve image details and generate more realistic images. HDM essentially is a DDPM that incorporates an attention mechanism between pixels. In HDM, the discrete form of the two-dimensional heat equation is integrated into the diffusion and generation formulas of DDPM, enabling the model to compute relationships between neighboring pixels during image processing. Our experiments demonstrate that HDM can generate higher-quality samples compared to models such as DDPM, Consistency Diffusion Models (CDM), Latent Diffusion Models (LDM), and Vector Quantized Generative Adversarial Networks (VQGAN).
Authors: Guanwen Feng, Zhiyuan Ma, Yunan Li, Jiahao Yang, Junwei Jing, Qiguang Miao
Abstract: Recent advances in audio-driven talking head generation have achieved impressive results in lip synchronization and emotional expression. However, they largely overlook the crucial task of facial attribute editing. This capability is indispensable for achieving deep personalization and expanding the range of practical applications, including user-tailored digital avatars, engaging online education content, and brand-specific digital customer service. In these key domains, flexible adjustment of visual attributes, such as hairstyle, accessories, and subtle facial features, is essential for aligning with user preferences, reflecting diverse brand identities and adapting to varying contextual demands. In this paper, we present FaceEditTalker, a unified framework that enables controllable facial attribute manipulation while generating high-quality, audio-synchronized talking head videos. Our method consists of two key components: an image feature space editing module, which extracts semantic and detail features and allows flexible control over attributes like expression, hairstyle, and accessories; and an audio-driven video generation module, which fuses these edited features with audio-guided facial landmarks to drive a diffusion-based generator. This design ensures temporal coherence, visual fidelity, and identity preservation across frames. Extensive experiments on public datasets demonstrate that our method achieves comparable or superior performance to representative baseline methods in lip-sync accuracy, video quality, and attribute controllability. Project page: https://peterfanfan.github.io/FaceEditTalker/
Authors: Lizhen Wang, Zhurong Xia, Tianshu Hu, Pengrui Wang, Pengfei Wei, Zerong Zheng, Ming Zhou, Yuan Zhang, Mingyuan Gao
Abstract: In e-commerce and digital marketing, generating high-fidelity human-product demonstration videos is important for effective product presentation. However, most existing frameworks either fail to preserve the identities of both humans and products or lack an understanding of human-product spatial relationships, leading to unrealistic representations and unnatural interactions. To address these challenges, we propose a Diffusion Transformer (DiT)-based framework. Our method simultaneously preserves human identities and product-specific details, such as logos and textures, by injecting paired human-product reference information and utilizing an additional masked cross-attention mechanism. We employ a 3D body mesh template and product bounding boxes to provide precise motion guidance, enabling intuitive alignment of hand gestures with product placements. Additionally, structured text encoding is used to incorporate category-level semantics, enhancing 3D consistency during small rotational changes across frames. Trained on a hybrid dataset with extensive data augmentation strategies, our approach outperforms state-of-the-art techniques in maintaining the identity integrity of both humans and products and generating realistic demonstration motions. Project page: https://lizhenwangt.github.io/DreamActor-H1/.
Authors: Xiaohui Jiang, Haijiang Zhu, Chade Li, Fulin Tang, Ning An
Abstract: LiDAR-based place recognition serves as a crucial enabler for long-term autonomy in robotics and autonomous driving systems. Yet, prevailing methodologies relying on handcrafted feature extraction face dual challenges: (1) Inconsistent point cloud density, induced by ego-motion dynamics and environmental disturbances during repeated traversals, leads to descriptor instability, and (2) Representation fragility stems from reliance on single-level geometric abstractions that lack discriminative power in structurally complex scenarios. To address these limitations, we propose a novel framework that redefines 3D place recognition through density-agnostic geometric reasoning. Specifically, we introduce an implicit 3D representation based on elastic points, which is immune to the interference of original scene point cloud density and achieves the characteristic of uniform distribution. Subsequently, we derive the occupancy grid and normal vector information of the scene from this implicit representation. Finally, with the aid of these two types of information, we obtain descriptors that fuse geometric information from both bird's-eye view (capturing macro-level spatial layouts) and 3D segment (encoding micro-scale surface geometries) perspectives. We conducted extensive experiments on numerous datasets (KITTI, KITTI-360, MulRan, NCLT) across diverse environments. The experimental results demonstrate that our method achieves state-of-the-art performance. Moreover, our approach strikes an optimal balance between accuracy, runtime, and memory optimization for historical maps, showcasing excellent Resilient and scalability. Our code will be open-sourced in the future.
Authors: Hangcheng Liu, Xu Kuang, Xingshuo Han, Xingwan Wu, Haoran Ou, Shangwei Guo, Xingyi Huang, Tao Xiang, Tianwei Zhang
Abstract: Stereo Depth Estimation (SDE) is essential for scene perception in vision-based systems such as autonomous driving. Prior work shows SDE is vulnerable to pixel-optimization attacks, but these methods are limited to digital, static, and view-specific settings, making them impractical. This raises a central question: how to design deployable, adaptive, and transferable attacks under realistic constraints? We present two contributions to answer it. First, we build a unified framework that extends pixel-optimization attacks to four stereo-matching stages: feature extraction, cost-volume construction, cost aggregation, and disparity regression. Through systematic evaluation across nine SDE models with realistic constraints like photometric consistency, we show existing attacks suffer from poor transferability. Second, we propose PatchHunter, the first pixel-optimization-free attack. PatchHunter casts patch generation as a search in a structured space of visual patterns that disrupt core SDE assumptions, and uses a reinforcement learning policy to discover effective and transferable patterns efficiently. We evaluate PatchHunter on three levels: autonomous driving dataset, high-fidelity simulator, and real-world deployment. On KITTI, PatchHunter outperforms pixel-level attacks in both effectiveness and black-box transferability. Tests in CARLA and on vehicles with industrial-grade stereo cameras confirm robustness to physical variations. Even under challenging conditions such as low lighting, PatchHunter achieves a D1-all error above 0.4, while pixel-level attacks remain near 0.
Authors: Jiaji Zhang, Ruichao Sun, Hailiang Zhao, Jiaju Wu, Peng Chen, Hao Li, Yuying Liu, Kingsum Chow, Gang Xiong, Shuiguang Deng
Abstract: Diffusion models have demonstrated exceptional generative capabilities but are computationally intensive, posing significant challenges for deployment in resource-constrained or latency-sensitive environments. Quantization offers an effective means to reduce model size and computational cost, with post-training quantization (PTQ) being particularly appealing due to its compatibility with pre-trained models without requiring retraining or training data. However, existing PTQ methods for diffusion models often rely on architecture-specific heuristics that limit their generalizability and hinder integration with industrial deployment pipelines. To address these limitations, we propose SegQuant, a unified quantization framework that adaptively combines complementary techniques to enhance cross-model versatility. SegQuant consists of a segment-aware, graph-based quantization strategy (SegLinear) that captures structural semantics and spatial heterogeneity, along with a dual-scale quantization scheme (DualScale) that preserves polarity-asymmetric activations, which is crucial for maintaining visual fidelity in generated outputs. SegQuant is broadly applicable beyond Transformer-based diffusion models, achieving strong performance while ensuring seamless compatibility with mainstream deployment tools.
Authors: Jijun Wang, Yan Wu, Yujian Mo, Junqiao Zhao, Jun Yan, Yinghao Hu
Abstract: Existing LiDAR-Camera fusion methods have achieved strong results in 3D object detection. To address the sparsity of point clouds, previous approaches typically construct spatial pseudo point clouds via depth completion as auxiliary input and adopts a proposal-refinement framework to generate detection results. However, introducing pseudo points inevitably brings noise, potentially resulting in inaccurate predictions. Considering the differing roles and reliability levels of each modality, we propose LDRFusion, a novel Lidar-dominant two-stage refinement framework for multi-sensor fusion. The first stage soley relies on LiDAR to produce accurately localized proposals, followed by a second stage where pseudo point clouds are incorporated to detect challenging instances. The instance-level results from both stages are subsequently merged. To further enhance the representation of local structures in pseudo point clouds, we present a hierarchical pseudo point residual encoding module, which encodes neighborhood sets using both feature and positional residuals. Experiments on the KITTI dataset demonstrate that our framework consistently achieves strong performance across multiple categories and difficulty levels.
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 Robust single-stage fully Sparse 3D object Detection Network 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: Shaoguang Wang, Ziyang Chen, Yijie Xu, 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: Yiyang Su, Yunping Shi, Feng Liu, Xiaoming Liu
Abstract: Recently, research interest in person re-identification (ReID) has increasingly focused on video-based scenarios, which are essential for robust surveillance and security in varied and dynamic environments. However, existing video-based ReID methods often overlook the necessity of identifying and selecting the most discriminative features from both videos in a query-gallery pair for effective matching. To address this issue, we propose a novel Hierarchical and Adaptive Mixture of Biometric Experts (HAMoBE) framework, which leverages multi-layer features from a pre-trained large model (e.g., CLIP) and is designed to mimic human perceptual mechanisms by independently modeling key biometric features--appearance, static body shape, and dynamic gait--and adaptively integrating them. Specifically, HAMoBE includes two levels: the first level extracts low-level features from multi-layer representations provided by the frozen large model, while the second level consists of specialized experts focusing on long-term, short-term, and temporal features. To ensure robust matching, we introduce a new dual-input decision gating network that dynamically adjusts the contributions of each expert based on their relevance to the input scenarios. Extensive evaluations on benchmarks like MEVID demonstrate that our approach yields significant performance improvements (e.g., +13.0% Rank-1 accuracy).
Authors: Tony Danjun Wang, Tobias Czempiel, Nassir Navab, Lennart Bastian
Abstract: Deep neural networks are prone to learning spurious correlations, exploiting dataset-specific artifacts rather than meaningful features for prediction. In surgical operating rooms (OR), these manifest through the standardization of smocks and gowns that obscure robust identifying landmarks, introducing model bias for tasks related to modeling OR personnel. Through gradient-based saliency analysis on two public OR datasets, we reveal that CNN models succumb to such shortcuts, fixating on incidental visual cues such as footwear beneath surgical gowns, distinctive eyewear, or other role-specific identifiers. Avoiding such biases is essential for the next generation of intelligent assistance systems in the OR, which should accurately recognize personalized workflow traits, such as surgical skill level or coordination with other staff members. We address this problem by encoding personnel as 3D point cloud sequences, disentangling identity-relevant shape and motion patterns from appearance-based confounders. Our experiments demonstrate that while RGB and geometric methods achieve comparable performance on datasets with apparent simulation artifacts, RGB models suffer a 12% accuracy drop in realistic clinical settings with decreased visual diversity due to standardizations. This performance gap confirms that geometric representations capture more meaningful biometric features, providing an avenue to developing robust methods of modeling humans in the OR.
Authors: Zheng Qin, Ruobing Zheng, Yabing Wang, Tianqi Li, Yi Yuan, Jingdong Chen, Le Wang
Abstract: While Multimodal Large Language Models (MLLMs) show immense promise for achieving truly human-like interactions, progress is hindered by the lack of fine-grained evaluation frameworks for human-centered scenarios, encompassing both the understanding of complex human intentions and the provision of empathetic, context-aware responses. Here we introduce HumanSense, a comprehensive benchmark designed to evaluate the human-centered perception and interaction capabilities of MLLMs, with a particular focus on deep understanding of extended multimodal contexts and the formulation of rational feedback. Our evaluation reveals that leading MLLMs still have considerable room for improvement, particularly for advanced interaction-oriented tasks. Supplementing visual input with audio and text information yields substantial improvements, and Omni-modal models show advantages on these tasks. Furthermore, we argue that appropriate feedback stems from a contextual analysis of the interlocutor's needs and emotions, with reasoning ability serving as the key to unlocking it. Accordingly, we employ a multi-stage, modality-progressive reinforcement learning to enhance the reasoning abilities of an Omni model, achieving substantial gains on evaluation results. Additionally, we observe that successful reasoning processes exhibit highly consistent thought patterns. By designing corresponding prompts, we also enhance the performance of non-reasoning models in a training-free manner. Project page: \textcolor{brightpink}https://digital-avatar.github.io/ai/HumanSense/
Authors: Ken Deng, Yunhan Yang, Jingxiang Sun, Xihui Liu, Yebin Liu, Ding Liang, Yan-Pei Cao
Abstract: We introduce GeoSAM2, a prompt-controllable framework for 3D part segmentation that casts the task as multi-view 2D mask prediction. Given a textureless object, we render normal and point maps from predefined viewpoints and accept simple 2D prompts - clicks or boxes - to guide part selection. These prompts are processed by a shared SAM2 backbone augmented with LoRA and residual geometry fusion, enabling view-specific reasoning while preserving pretrained priors. The predicted masks are back-projected to the object and aggregated across views. Our method enables fine-grained, part-specific control without requiring text prompts, per-shape optimization, or full 3D labels. In contrast to global clustering or scale-based methods, prompts are explicit, spatially grounded, and interpretable. We achieve state-of-the-art class-agnostic performance on PartObjaverse-Tiny and PartNetE, outperforming both slow optimization-based pipelines and fast but coarse feedforward approaches. Our results highlight a new paradigm: aligning the paradigm of 3D segmentation with SAM2, leveraging interactive 2D inputs to unlock controllability and precision in object-level part understanding.
Authors: Darya Taratynova, Alya Almsouti, Beknur Kalmakhanbet, Numan Saeed, Mohammad Yaqub
Abstract: Congenital heart defect (CHD) detection in ultrasound videos is hindered by image noise and probe positioning variability. While automated methods can reduce operator dependence, current machine learning approaches often neglect temporal information, limit themselves to binary classification, and do not account for prediction calibration. We propose Temporal Prompt Alignment (TPA), a method leveraging foundation image-text model and prompt-aware contrastive learning to classify fetal CHD on cardiac ultrasound videos. TPA extracts features from each frame of video subclips using an image encoder, aggregates them with a trainable temporal extractor to capture heart motion, and aligns the video representation with class-specific text prompts via a margin-hinge contrastive loss. To enhance calibration for clinical reliability, we introduce a Conditional Variational Autoencoder Style Modulation (CVAESM) module, which learns a latent style vector to modulate embeddings and quantifies classification uncertainty. Evaluated on a private dataset for CHD detection and on a large public dataset, EchoNet-Dynamic, for systolic dysfunction, TPA achieves state-of-the-art macro F1 scores of 85.40% for CHD diagnosis, while also reducing expected calibration error by 5.38% and adaptive ECE by 6.8%. On EchoNet-Dynamic's three-class task, it boosts macro F1 by 4.73% (from 53.89% to 58.62%). Temporal Prompt Alignment (TPA) is a framework for fetal congenital heart defect (CHD) classification in ultrasound videos that integrates temporal modeling, prompt-aware contrastive learning, and uncertainty quantification.
Authors: Naen Xu, Jinghuai Zhang, Changjiang Li, Zhi Chen, Chunyi Zhou, Qingming Li, Tianyu Du, Shouling Ji
Abstract: The rapid growth of text-to-video (T2V) diffusion models has raised concerns about privacy, copyright, and safety due to their potential misuse in generating harmful or misleading content. These models are often trained on numerous datasets, including unauthorized personal identities, artistic creations, and harmful materials, which can lead to uncontrolled production and distribution of such content. To address this, we propose VideoEraser, a training-free framework that prevents T2V diffusion models from generating videos with undesirable concepts, even when explicitly prompted with those concepts. Designed as a plug-and-play module, VideoEraser can seamlessly integrate with representative T2V diffusion models via a two-stage process: Selective Prompt Embedding Adjustment (SPEA) and Adversarial-Resilient Noise Guidance (ARNG). We conduct extensive evaluations across four tasks, including object erasure, artistic style erasure, celebrity erasure, and explicit content erasure. Experimental results show that VideoEraser consistently outperforms prior methods regarding efficacy, integrity, fidelity, robustness, and generalizability. Notably, VideoEraser achieves state-of-the-art performance in suppressing undesirable content during T2V generation, reducing it by 46% on average across four tasks compared to baselines.
Authors: Ruizhuo Song, Beiming Yuan
Abstract: Despite the outstanding performance of current deep learning models across various domains, their fundamental bottleneck in abstract reasoning remains unresolved. To address this challenge, the academic community has introduced Raven's Progressive Matrices (RPM) problems as an authoritative benchmark for evaluating the abstract reasoning capabilities of deep learning algorithms, with a focus on core intelligence dimensions such as abstract reasoning, pattern recognition, and complex problem-solving. Therefore, this paper centers on solving RPM problems, aiming to contribute to enhancing the abstract reasoning abilities of machine intelligence. Firstly, this paper adopts a ``causal chain modeling'' perspective to systematically analyze the complete causal chain in RPM tasks: image $\rightarrow$ abstract attributes $\rightarrow$ progressive attribute patterns $\rightarrow$ pattern consistency $\rightarrow$ correct answer. Based on this analysis, the network architecture of the baseline model DIO is designed. However, experiments reveal that the optimization objective formulated for DIO, namely maximizing the variational lower bound of mutual information between the context and the correct option, fails to enable the model to genuinely acquire the predefined human reasoning logic. This is attributed to two main reasons: the tightness of the lower bound significantly impacts the effectiveness of mutual information maximization, and mutual information, as a statistical measure, does not capture the causal relationship between subjects and objects. To overcome these limitations, this paper progressively proposes three improvement methods:
Authors: Ashwath Vaithinathan Aravindan, Abha Jha, Mihir Kulkarni
Abstract: Vision-Language Models (VLMs) have shown remarkable performance in integrating visual and textual information for tasks such as image captioning and visual question answering. However, these models struggle with compositional generalization and object binding, which limit their ability to handle novel combinations of objects and their attributes. Our work explores the root causes of these failures using mechanistic interpretability techniques. We show evidence that individual neurons in the MLP layers of CLIP's vision encoder represent multiple features, and this "superposition" directly hinders its compositional feature representation which consequently affects compositional reasoning and object binding capabilities. We hope this study will serve as an initial step toward uncovering the mechanistic roots of compositional failures in VLMs. The code and supporting results can be found https://github.com/Mystic-Slice/Do-VLMs-Have-Bad-Eyes.
URLs: https://github.com/Mystic-Slice/Do-VLMs-Have-Bad-Eyes.
Authors: Fucai Ke, Joy Hsu, Zhixi Cai, Zixian Ma, Xin Zheng, Xindi Wu, Sukai Huang, Weiqing Wang, Pari Delir Haghighi, Gholamreza Haffari, Ranjay Krishna, Jiajun Wu, Hamid Rezatofighi
Abstract: Compositional visual reasoning has emerged as a key research frontier in multimodal AI, aiming to endow machines with the human-like ability to decompose visual scenes, ground intermediate concepts, and perform multi-step logical inference. While early surveys focus on monolithic vision-language models or general multimodal reasoning, a dedicated synthesis of the rapidly expanding compositional visual reasoning literature is still missing. We fill this gap with a comprehensive survey spanning 2023 to 2025 that systematically reviews 260+ papers from top venues (CVPR, ICCV, NeurIPS, ICML, ACL, etc.). We first formalize core definitions and describe why compositional approaches offer advantages in cognitive alignment, semantic fidelity, robustness, interpretability, and data efficiency. Next, we trace a five-stage paradigm shift: from prompt-enhanced language-centric pipelines, through tool-enhanced LLMs and tool-enhanced VLMs, to recently minted chain-of-thought reasoning and unified agentic VLMs, highlighting their architectural designs, strengths, and limitations. We then catalog 60+ benchmarks and corresponding metrics that probe compositional visual reasoning along dimensions such as grounding accuracy, chain-of-thought faithfulness, and high-resolution perception. Drawing on these analyses, we distill key insights, identify open challenges (e.g., limitations of LLM-based reasoning, hallucination, a bias toward deductive reasoning, scalable supervision, tool integration, and benchmark limitations), and outline future directions, including world-model integration, human-AI collaborative reasoning, and richer evaluation protocols. By offering a unified taxonomy, historical roadmap, and critical outlook, this survey aims to serve as a foundational reference and inspire the next generation of compositional visual reasoning research.
Authors: Weiyun Wang, Zhangwei Gao, Lixin Gu, Hengjun Pu, Long Cui, Xingguang Wei, Zhaoyang Liu, Linglin Jing, Shenglong Ye, Jie Shao, Zhaokai Wang, Zhe Chen, Hongjie Zhang, Ganlin Yang, Haomin Wang, Qi Wei, Jinhui Yin, Wenhao Li, Erfei Cui, Guanzhou Chen, Zichen Ding, Changyao Tian, Zhenyu Wu, Jingjing Xie, Zehao Li, Bowen Yang, Yuchen Duan, Xuehui Wang, Zhi Hou, Haoran Hao, Tianyi Zhang, Songze Li, Xiangyu Zhao, Haodong Duan, Nianchen Deng, Bin Fu, Yinan He, Yi Wang, Conghui He, Botian Shi, Junjun He, Yingtong Xiong, Han Lv, Lijun Wu, Wenqi Shao, Kaipeng Zhang, Huipeng Deng, Biqing Qi, Jiaye Ge, Qipeng Guo, Wenwei Zhang, Songyang Zhang, Maosong Cao, Junyao Lin, Kexian Tang, Jianfei Gao, Haian Huang, Yuzhe Gu, Chengqi Lyu, Huanze Tang, Rui Wang, Haijun Lv, Wanli Ouyang, Limin Wang, Min Dou, Xizhou Zhu, Tong Lu, Dahua Lin, Jifeng Dai, Weijie Su, Bowen Zhou, Kai Chen, Yu Qiao, Wenhai Wang, Gen Luo
Abstract: We introduce InternVL 3.5, a new family of open-source multimodal models that significantly advances versatility, reasoning capability, and inference efficiency along the InternVL series. A key innovation is the Cascade Reinforcement Learning (Cascade RL) framework, which enhances reasoning through a two-stage process: offline RL for stable convergence and online RL for refined alignment. This coarse-to-fine training strategy leads to substantial improvements on downstream reasoning tasks, e.g., MMMU and MathVista. To optimize efficiency, we propose a Visual Resolution Router (ViR) that dynamically adjusts the resolution of visual tokens without compromising performance. Coupled with ViR, our Decoupled Vision-Language Deployment (DvD) strategy separates the vision encoder and language model across different GPUs, effectively balancing computational load. These contributions collectively enable InternVL3.5 to achieve up to a +16.0\% gain in overall reasoning performance and a 4.05$\times$ inference speedup compared to its predecessor, i.e., InternVL3. In addition, InternVL3.5 supports novel capabilities such as GUI interaction and embodied agency. Notably, our largest model, i.e., InternVL3.5-241B-A28B, attains state-of-the-art results among open-source MLLMs across general multimodal, reasoning, text, and agentic tasks -- narrowing the performance gap with leading commercial models like GPT-5. All models and code are publicly released.
Authors: Md. Rashid Shahriar Khan, Md. Abrar Hasan, Mohammod Tareq Aziz Justice
Abstract: Detecting anomalies in surveillance footage is inherently challenging due to their unpredictable and context-dependent nature. This work introduces a novel context-aware zero-shot anomaly detection framework that identifies abnormal events without exposure to anomaly examples during training. The proposed hybrid architecture combines TimeSformer, DPC, and CLIP to model spatiotemporal dynamics and semantic context. TimeSformer serves as the vision backbone to extract rich spatial-temporal features, while DPC forecasts future representations to identify temporal deviations. Furthermore, a CLIP-based semantic stream enables concept-level anomaly detection through context-specific text prompts. These components are jointly trained using InfoNCE and CPC losses, aligning visual inputs with their temporal and semantic representations. A context-gating mechanism further enhances decision-making by modulating predictions with scene-aware cues or global video features. By integrating predictive modeling with vision-language understanding, the system can generalize to previously unseen behaviors in complex environments. This framework bridges the gap between temporal reasoning and semantic context in zero-shot anomaly detection for surveillance. The code for this research has been made available at https://github.com/NK-II/Context-Aware-Zero-Shot-Anomaly-Detection-in-Surveillance.
URLs: https://github.com/NK-II/Context-Aware-Zero-Shot-Anomaly-Detection-in-Surveillance.
Authors: Chunlin Zhong, Qiuxia Hou, Zhangjun Zhou, Shuang Hao, Haonan Lu, Yanhao Zhang, He Tang, Xiang Bai
Abstract: Video captioning aims to generate comprehensive and coherent descriptions of the video content, contributing to the advancement of both video understanding and generation. However, existing methods often suffer from motion-detail imbalance, as models tend to overemphasize one aspect while neglecting the other. This imbalance results in incomplete captions, which in turn leads to a lack of consistency in video understanding and generation. To address this issue, we propose solutions from two aspects: 1) Data aspect: We constructed the Harmonizing Motion-Detail 270K (HMD-270K) dataset through a two-stage pipeline: Motion-Detail Fusion (MDF) and Fine-Grained Examination (FGE). 2) Optimization aspect: We introduce the Caption Set Equivalence Reward (CSER) based on Group Relative Policy Optimization (GRPO). CSER enhances completeness and accuracy in capturing both motion and details through unit-to-set matching and bidirectional validation. Based on the HMD-270K supervised fine-tuning and GRPO post-training with CSER, we developed OwlCap, a powerful video captioning multi-modal large language model (MLLM) with motion-detail balance. Experimental results demonstrate that OwlCap achieves significant improvements compared to baseline models on two benchmarks: the detail-focused VDC (+4.2 Acc) and the motion-focused DREAM-1K (+4.6 F1). The HMD-270K dataset and OwlCap model will be publicly released to facilitate video captioning research community advancements.
Authors: Jeonghwan Kim, Yushi Lan, Armando Fortes, Yongwei Chen, Xingang Pan
Abstract: Recent mesh generation approaches typically tokenize triangle meshes into sequences of tokens and train autoregressive models to generate these tokens sequentially. Despite substantial progress, such token sequences inevitably reuse vertices multiple times to fully represent manifold meshes, as each vertex is shared by multiple faces. This redundancy leads to excessively long token sequences and inefficient generation processes. In this paper, we propose an efficient framework that generates artistic meshes by treating vertices and faces separately, significantly reducing redundancy. We employ an autoregressive model solely for vertex generation, decreasing the token count to approximately 23\% of that required by the most compact existing tokenizer. Next, we leverage a bidirectional transformer to complete the mesh in a single step by capturing inter-vertex relationships and constructing the adjacency matrix that defines the mesh faces. To further improve the generation quality, we introduce a fidelity enhancer to refine vertex positioning into more natural arrangements and propose a post-processing framework to remove undesirable edge connections. Experimental results show that our method achieves more than 8$\times$ faster speed on mesh generation compared to state-of-the-art approaches, while producing higher mesh quality.
Authors: Xiang Li, Wenyue Hua, Kaijie Zhu, Lingyao Li, Haoyang Ling, Jinkui Chi, Qi Dou, Jindong Wang, Yongfeng Zhang, Xin Ma, Lizhou Fan
Abstract: Large Vision-Language Models (LVLMs) have demonstrated impressive capabilities in multimodal understanding, yet their reasoning abilities remain underexplored. Existing benchmarks tend to focus on perception or text-based comprehension, offering limited insight into how well these models perform on structured, logic-driven tasks that require both visual and linguistic reasoning. To address this gap, we introduce NPHardEval4V, a multimodal benchmark suite grounded in four classical NP-hard problems: Knapsack, Set Cover, Traveling Salesperson, and Vertex Cover. Each task is presented through a combination of structured visual layouts and textual prompts, designed to assess the ability of LVLMs to perform combinatorial reasoning under visual-linguistic constraints. We evaluate a set of advanced open-source and closed-source vision-language models under a unified prompting and problem representation framework. This enables fair comparison across models and task types, while isolating key variables affecting performance. Our results show that while these models perform reasonably well on perception-based inputs, they struggle with global optimization, abstraction, and constraint satisfaction. No single model demonstrates consistent reasoning capability across all problem types, and common failure patterns reveal fundamental limitations in current architectures. By leveraging the structure and complexity of NP-hard problems, NPHardEval4V provides a scalable, interpretable, and challenging testbed for diagnosing reasoning behaviors in LVLMs. We hope this benchmark can support the community in building more robust, inference-capable multimodal systems. The benchmark dataset and code are available at https://github.com/lizhouf/NPHardEval4.
Authors: Yixin Huang, Yiqi Jin, Ke Tao, Kaijian Xia, Jianfeng Gu, Lei Yu, Haojie Li, Lan Du, Cunjian Chen
Abstract: May-Thurner Syndrome (MTS) is a vascular condition that affects over 20\% of the population and significantly increases the risk of iliofemoral deep venous thrombosis. Accurate and early diagnosis of MTS using computed tomography (CT) remains a clinical challenge due to the subtle anatomical compression and variability across patients. In this paper, we propose MTS-Net, an end-to-end 3D deep learning framework designed to capture spatial-temporal patterns from CT volumes for reliable MTS diagnosis. MTS-Net builds upon 3D ResNet-18 by embedding a novel dual-enhanced positional multi-head self-attention (DEP-MHSA) module into the Transformer encoder of the network's final stages. The proposed DEP-MHSA employs multi-scale convolution and integrates positional embeddings into both attention weights and residual paths, enhancing spatial context preservation, which is crucial for identifying venous compression. To validate our approach, we curate the first publicly available dataset for MTS, MTS-CT, containing over 747 gender-balanced subjects with standard and enhanced CT scans. Experimental results demonstrate that MTS-Net achieves average 0.79 accuracy, 0.84 AUC, and 0.78 F1-score, outperforming baseline models including 3D ResNet, DenseNet-BC, and BabyNet. Our work not only introduces a new diagnostic architecture for MTS but also provides a high-quality benchmark dataset to facilitate future research in automated vascular syndrome detection. We make our code and dataset publicly available at:https://github.com/Nutingnon/MTS_dep_mhsa.
Authors: Ethan Kane Waters, Carla Chia-ming Chen, Mostafa Rahimi Azghadi
Abstract: Disease detection in sugarcane, particularly the identification of asymptomatic infectious diseases such as Ratoon Stunting Disease (RSD), is critical for effective crop management. This study employed various machine learning techniques to detect the presence of RSD in different sugarcane varieties, using vegetation indices derived from freely available satellite-based spectral data. Our results show that the Support Vector Machine with a Radial Basis Function Kernel (SVM-RBF) was the most effective algorithm, achieving classification accuracy between 85.64% and 96.55%, depending on the variety. Gradient Boosting and Random Forest also demonstrated high performance achieving accuracy between 83.33% to 96.55%, while Logistic Regression and Quadratic Discriminant Analysis showed variable results across different varieties. The inclusion of sugarcane variety and vegetation indices was important in the detection of RSD. This agreed with what was identified in the current literature. Our study highlights the potential of satellite-based remote sensing as a cost-effective and efficient method for large-scale sugarcane disease detection alternative to traditional manual laboratory testing methods.
Authors: Matthieu Kowalski, Beno\^it Mal\'ezieux, Thomas Moreau, Audrey Repetti
Abstract: In this work we study the behavior of the forward-backward (FB) algorithm when the proximity operator is replaced by a sub-iterative procedure to approximate a Gaussian denoiser, in a Plug-and-Play (PnP) fashion. In particular, we consider both analysis and synthesis Gaussian denoisers within a dictionary framework, obtained by unrolling dual-FB iterations or FB iterations, respectively. We analyze the associated minimization problems as well as the asymptotic behavior of the resulting FB-PnP iterations. In particular, we show that the synthesis Gaussian denoising problem can be viewed as a proximity operator. For each case, analysis and synthesis, we show that the FB-PnP algorithms solve the same problem whether we use only one or an infinite number of sub-iteration to solve the denoising problem at each iteration. To this aim, we show that each "one sub-iteration" strategy within the FB-PnP can be interpreted as a primal-dual algorithm when a warm-restart strategy is used. We further present similar results when using a Moreau-Yosida smoothing of the global problem, for an arbitrary number of sub-iterations. Finally, we provide numerical simulations to illustrate our theoretical results. In particular we first consider a toy compressive sensing example, as well as an image restoration problem in a deep dictionary framework.
Authors: Daixuan Cheng, Shaohan Huang, Ziyu Zhu, Xintong Zhang, Wayne Xin Zhao, Zhongzhi Luan, Bo Dai, Zhenliang Zhang
Abstract: Adapting general multimodal large language models (MLLMs) to specific domains, such as scientific and industrial fields, is highly significant in promoting their practical applications. This paper systematically investigates domain adaptation of MLLMs via post-training, focusing on data synthesis, training pipeline, and task evaluation. (1) Data Synthesis: Using only open-source models, we develop a generate-then-filter pipeline that curates diverse visual instruction tasks based on domain-specific image-caption pairs. The resulting data surpass the data synthesized by manual rules or strong closed-source models in enhancing domain-specific performance. (2) Training Pipeline: Unlike general MLLMs that typically adopt a two-stage training paradigm, we find that a single-stage approach is more effective for domain adaptation. (3) Task Evaluation: We conduct extensive experiments in high-impact domains such as biomedicine, food, and remote sensing, by post-training a variety of MLLMs and then evaluating MLLM performance on various domain-specific tasks. Finally, we fully open-source our models, code, and data to encourage future research in this area.
Authors: Joshua Bird, Jan Blumenkamp, Amanda Prorok
Abstract: Cooperative Simultaneous Localization and Mapping (C-SLAM) enables multiple agents to work together in mapping unknown environments while simultaneously estimating their own positions. This approach enhances robustness, scalability, and accuracy by sharing information between agents, reducing drift, and enabling collective exploration of larger areas. In this paper, we present Decentralized Visual Monocular SLAM (DVM-SLAM), the first open-source decentralized monocular C-SLAM system. By only utilizing low-cost and light-weight monocular vision sensors, our system is well suited for small robots and micro aerial vehicles (MAVs). DVM-SLAM's real-world applicability is validated on physical robots with a custom collision avoidance framework, showcasing its potential in real-time multi-agent autonomous navigation scenarios. We also demonstrate comparable accuracy to state-of-the-art centralized monocular C-SLAM systems. We open-source our code and provide supplementary material online.
Authors: Yanfei Li, Teng Yin, Wenyi Shang, Jingyu Liu, Xi Wang, Kaiyang Zhao
Abstract: Missing modalities pose a major issue in Alzheimer's Disease (AD) diagnosis, as many subjects lack full imaging data due to cost and clinical constraints. While multi-modal learning leverages complementary information, most existing methods train only on complete data, ignoring the large proportion of incomplete samples in real-world datasets like ADNI. This reduces the effective training set and limits the full use of valuable medical data. While some methods incorporate incomplete samples, they fail to effectively address inter-modal feature alignment and knowledge transfer challenges under high missing rates. To address this, we propose a Prototype-Guided Adaptive Distillation (PGAD) framework that directly incorporates incomplete multi-modal data into training. PGAD enhances missing modality representations through prototype matching and balances learning with a dynamic sampling strategy. We validate PGAD on the ADNI dataset with varying missing rates (20%, 50%, and 70%) and demonstrate that it significantly outperforms state-of-the-art approaches. Ablation studies confirm the effectiveness of prototype matching and adaptive sampling, highlighting the potential of our framework for robust and scalable AD diagnosis in real-world clinical settings.
Authors: Lijun Sheng, Jian Liang, Zilei Wang, Ran He
Abstract: Vision-language models (VLMs), such as CLIP, have gained significant popularity as foundation models, with numerous fine-tuning methods developed to enhance performance on downstream tasks. However, due to their inherent vulnerability and the common practice of selecting from a limited set of open-source models, VLMs suffer from a higher risk of adversarial attacks than traditional vision models. Existing defense techniques typically rely on adversarial fine-tuning during training, which requires labeled data and lacks of flexibility for downstream tasks. To address these limitations, we propose robust test-time prompt tuning (R-TPT), which mitigates the impact of adversarial attacks during the inference stage. We first reformulate the classic marginal entropy objective by eliminating the term that introduces conflicts under adversarial conditions, retaining only the pointwise entropy minimization. Furthermore, we introduce a plug-and-play reliability-based weighted ensembling strategy, which aggregates useful information from reliable augmented views to strengthen the defense. R-TPT enhances defense against adversarial attacks without requiring labeled training data while offering high flexibility for inference tasks. Extensive experiments on widely used benchmarks with various attacks demonstrate the effectiveness of R-TPT. The code is available in https://github.com/TomSheng21/R-TPT.
Authors: Xilong Wang, John Bloch, Zedian Shao, Yuepeng Hu, Shuyan Zhou, Neil Zhenqiang Gong
Abstract: Multi-modal large language model (MLLM)-based web agents interact with webpage environments by generating actions based on screenshots of the webpages. Environmental prompt injection attacks manipulate the environment to induce the web agent to perform a specific, attacker-chosen action--denoted as the target action. However, existing attacks suffer from limited effectiveness or stealthiness, or are impractical in real-world settings. In this work, we propose EnvInjection, a new attack that addresses these limitations. Our attack adds a perturbation to the raw pixel values of the rendered webpage. After these perturbed pixels are mapped into a screenshot, the perturbation induces the web agent to perform the target action. We formulate the task of finding the perturbation as an optimization problem. A key challenge in solving this problem is that the mapping between raw pixel values and screenshot is non-differentiable, making it difficult to backpropagate gradients to the perturbation. To overcome this, we train a neural network to approximate the mapping and apply projected gradient descent to solve the reformulated optimization problem. Extensive evaluation on multiple webpage datasets shows that EnvInjection is highly effective and significantly outperforms existing baselines.
Authors: Sirui Li, Linkai Peng, Zheyuan Zhang, Gorkem Durak, Ulas Bagci
Abstract: Foundation models (FMs) such as CLIP and SAM have recently shown great promise in image segmentation tasks, yet their adaptation to 3D medical imaging-particularly for pathology detection and segmentation-remains underexplored. A critical challenge arises from the domain gap between natural images and medical volumes: existing FMs, pre-trained on 2D data, struggle to capture 3D anatomical context, limiting their utility in clinical applications like tumor segmentation. To address this, we propose an adaptation framework called TAGS: Tumor Adaptive Guidance for SAM, which unlocks 2D FMs for 3D medical tasks through multi-prompt fusion. By preserving most of the pre-trained weights, our approach enhances SAM's spatial feature extraction using CLIP's semantic insights and anatomy-specific prompts. Extensive experiments on three open-source tumor segmentation datasets prove that our model surpasses the state-of-the-art medical image segmentation models (+46.88% over nnUNet), interactive segmentation frameworks, and other established medical FMs, including SAM-Med2D, SAM-Med3D, SegVol, Universal, 3D-Adapter, and SAM-B (at least +13% over them). This highlights the robustness and adaptability of our proposed framework across diverse medical segmentation tasks.
Authors: George Webber, Alexander Hammers, Andrew P. King, Andrew J. Reader
Abstract: Recent work has shown improved lesion detectability and flexibility to reconstruction hyperparameters (e.g. scanner geometry or dose level) when PET images are reconstructed by leveraging pre-trained diffusion models. Such methods train a diffusion model (without sinogram data) on high-quality, but still noisy, PET images. In this work, we propose a simple method for generating subject-specific PET images from a dataset of multi-subject PET-MR scans, synthesizing "pseudo-PET" images by transforming between different patients' anatomy using image registration. The images we synthesize retain information from the subject's MR scan, leading to higher resolution and the retention of anatomical features compared to the original set of PET images. With simulated and real [$^{18}$F]FDG datasets, we show that pre-training a personalized diffusion model with subject-specific "pseudo-PET" images improves reconstruction accuracy with low-count data. In particular, the method shows promise in combining information from a guidance MR scan without overly imposing anatomical features, demonstrating an improved trade-off between reconstructing PET-unique image features versus features present in both PET and MR. We believe this approach for generating and utilizing synthetic data has further applications to medical imaging tasks, particularly because patient-specific PET images can be generated without resorting to generative deep learning or large training datasets.
Authors: Wei Cao, Marcel Hallgarten, Tianyu Li, Daniel Dauner, Xunjiang Gu, Caojun Wang, Yakov Miron, Marco Aiello, Hongyang Li, Igor Gilitschenski, Boris Ivanovic, Marco Pavone, Andreas Geiger, Kashyap Chitta
Abstract: Existing evaluation paradigms for Autonomous Vehicles (AVs) face critical limitations. Real-world evaluation is often challenging due to safety concerns and a lack of reproducibility, whereas closed-loop simulation can face insufficient realism or high computational costs. Open-loop evaluation, while being efficient and data-driven, relies on metrics that generally overlook compounding errors. In this paper, we propose pseudo-simulation, a novel paradigm that addresses these limitations. Pseudo-simulation operates on real datasets, similar to open-loop evaluation, but augments them with synthetic observations generated prior to evaluation using 3D Gaussian Splatting. Our key idea is to approximate potential future states the AV might encounter by generating a diverse set of observations that vary in position, heading, and speed. Our method then assigns a higher importance to synthetic observations that best match the AV's likely behavior using a novel proximity-based weighting scheme. This enables evaluating error recovery and the mitigation of causal confusion, as in closed-loop benchmarks, without requiring sequential interactive simulation. We show that pseudo-simulation is better correlated with closed-loop simulations ($R^2=0.8$) than the best existing open-loop approach ($R^2=0.7$). We also establish a public leaderboard for the community to benchmark new methodologies with pseudo-simulation. Our code is available at https://github.com/autonomousvision/navsim.
Authors: Tianxing Chen, Zanxin Chen, Baijun Chen, Zijian Cai, Yibin Liu, Zixuan Li, Qiwei Liang, Xianliang Lin, Yiheng Ge, Zhenyu Gu, Weiliang Deng, Yubin Guo, Tian Nian, Xuanbing Xie, Qiangyu Chen, Kailun Su, Tianling Xu, Guodong Liu, Mengkang Hu, Huan-ang Gao, Kaixuan Wang, Zhixuan Liang, Yusen Qin, Xiaokang Yang, Ping Luo, Yao Mu
Abstract: Simulation-based data synthesis has emerged as a powerful paradigm for advancing real-world robotic manipulation. Yet existing datasets remain insufficient for robust bimanual manipulation due to (1) the lack of scalable task generation methods and (2) oversimplified simulation environments. We present RoboTwin 2.0, a scalable framework for automated, large-scale generation of diverse and realistic data, together with unified evaluation protocols for dual-arm manipulation. At its core is RoboTwin-OD, an object library of 731 instances across 147 categories with semantic and manipulation-relevant annotations. Building on this, we design an expert data synthesis pipeline that leverages multimodal language models (MLLMs) and simulation-in-the-loop refinement to automatically generate task-level execution code. To improve sim-to-real transfer, RoboTwin 2.0 applies structured domain randomization along five axes: clutter, lighting, background, tabletop height, and language, enhancing data diversity and policy robustness. The framework is instantiated across 50 dual-arm tasks and five robot embodiments. Empirically, it yields a 10.9% gain in code generation success rate. For downstream policy learning, a VLA model trained with synthetic data plus only 10 real demonstrations achieves a 367% relative improvement over the 10-demo baseline, while zero-shot models trained solely on synthetic data obtain a 228% gain. These results highlight the effectiveness of RoboTwin 2.0 in strengthening sim-to-real transfer and robustness to environmental variations. We release the data generator, benchmark, dataset, and code to support scalable research in robust bimanual manipulation. Project Page: https://robotwin-platform.github.io/, Code: https://github.com/robotwin-Platform/robotwin/.
URLs: https://robotwin-platform.github.io/,, https://github.com/robotwin-Platform/robotwin/.
Authors: Tom\'as Silva Santos Rocha, Anastasiia Mikhailova, Moreno I. Coco, Jos\'e Santos-Victor
Abstract: The global prevalence of dementia is projected to double by 2050, highlighting the urgent need for scalable diagnostic tools. This study utilizes digital cognitive tasks with eye-tracking data correlated with memory processes to distinguish between Healthy Controls (HC) and Mild Cognitive Impairment (MCI), a precursor to dementia. A deep learning model based on VTNet was trained using eye-tracking data from 44 participants (24 MCI, 20 HCs) who performed a visual memory task. The model utilizes both time series and spatial data derived from eye-tracking. It was modified to incorporate scan paths, heat maps, and image content. These modifications also enabled testing parameters such as image resolution and task performance, analyzing their impact on model performance. The best model, utilizing $700\times700px$ resolution heatmaps, achieved 68% sensitivity and 76% specificity. Despite operating under more challenging conditions (e.g., smaller dataset size, shorter task duration, or a less standardized task), the model's performance is comparable to an Alzheimer's study using similar methods (70% sensitivity and 73% specificity). These findings contribute to the development of automated diagnostic tools for MCI. Future work should focus on refining the model and using a standardized long-term visual memory task.
Authors: Shitian Zhao, Haoquan Zhang, Shaoheng Lin, Ming Li, Qilong Wu, Kaipeng Zhang, Chen Wei
Abstract: LLMs are increasingly deployed as agents, systems capable of planning, reasoning, and dynamically calling external tools. However, in visual reasoning, prior approaches largely remain limited by predefined workflows and static toolsets. In this report, we present PyVision, an interactive, multi-turn framework that enables MLLMs to autonomously generate, execute, and refine Python-based tools tailored to the task at hand, unlocking flexible and interpretable problem-solving. We develop a taxonomy of the tools created by PyVision and analyze their usage across a diverse set of benchmarks. Quantitatively, PyVision achieves consistent performance gains, boosting GPT-4.1 by +7.8% on V* and Claude-4.0-Sonnet by +31.1% on VLMsAreBlind-mini. These results point to a broader shift: dynamic tooling allows models not just to use tools, but to invent them, advancing toward more agentic visual reasoning.
Authors: Rushin H. Gindra, Giovanni Palla, Mathias Nguyen, Sophia J. Wagner, Manuel Tran, Fabian J Theis, Dieter Saur, Lorin Crawford, Tingying Peng
Abstract: Spatial transcriptomics enables simultaneous measurement of gene expression and tissue morphology, offering unprecedented insights into cellular organization and disease mechanisms. However, the field lacks comprehensive benchmarks for evaluating multimodal learning methods that leverage both histology images and gene expression data. Here, we present HESCAPE, a large-scale benchmark for cross-modal contrastive pretraining in spatial transcriptomics, built on a curated pan-organ dataset spanning 6 different gene panels and 54 donors. We systematically evaluated state-of-the-art image and gene expression encoders across multiple pretraining strategies and assessed their effectiveness on two downstream tasks: gene mutation classification and gene expression prediction. Our benchmark demonstrates that gene expression encoders are the primary determinant of strong representational alignment, and that gene models pretrained on spatial transcriptomics data outperform both those trained without spatial data and simple baseline approaches. However, downstream task evaluation reveals a striking contradiction: while contrastive pretraining consistently improves gene mutation classification performance, it degrades direct gene expression prediction compared to baseline encoders trained without cross-modal objectives. We identify batch effects as a key factor that interferes with effective cross-modal alignment. Our findings highlight the critical need for batch-robust multimodal learning approaches in spatial transcriptomics. To accelerate progress in this direction, we release HESCAPE, providing standardized datasets, evaluation protocols, and benchmarking tools for the community
Authors: Prateek Poudel, Prashant Aryal, Kirtan Kunwar, Navin Nepal, Dinesh Baniya Kshatri
Abstract: We investigate the use of wavelet-space feature decomposition in neural super-resolution for rendering pipelines. Building on neural upscaling frameworks, we introduce a wavelet-domain representation that separates low-frequency and high-frequency details before reconstruction, enabling the network to better preserve fine textures while maintaining structural consistency. Unlike RGB-space regression, our approach leverages the stationary wavelet transform (SWT) to avoid spatial down-sampling, ensuring alignment across subbands and preserving shift invariance. The model predicts wavelet coefficients conditioned on spatial G-buffers and temporally warped history frames, which are then recombined through inverse wavelet synthesis. We conduct a comprehensive ablation study across wavelet families, transform types, and architectural variants, showing that incorporating SWT improves PSNR by 1.5 dB and reduces LPIPS by 17% on average, with only a modest relative runtime overhead. Taken together, our results suggest that wavelet-domain representations a principled path toward higher-quality super-resolution in graphics applications.
Authors: H\'el\`ene Corbaz, Anh Nguyen, Victor Schulze-Zachau, Paul Friedrich, Alicia Durrer, Florentin Bieder, Philippe C. Cattin, Marios N Psychogios
Abstract: Patients undergoing a mechanical thrombectomy procedure usually have a multi-detector CT (MDCT) scan before and after the intervention. The image quality of the flat panel detector CT (FDCT) present in the intervention room is generally much lower than that of a MDCT due to significant artifacts. However, using only FDCT images could improve patient management as the patient would not need to be moved to the MDCT room. Several studies have evaluated the potential use of FDCT imaging alone and the time that could be saved by acquiring the images before and/or after the intervention only with the FDCT. This study proposes using a denoising diffusion probabilistic model (DDPM) to improve the image quality of FDCT scans, making them comparable to MDCT scans. Clinicans evaluated FDCT, MDCT, and our model's predictions for diagnostic purposes using a questionnaire. The DDPM eliminated most artifacts and improved anatomical visibility without reducing bleeding detection, provided that the input FDCT image quality is not too low. Our code can be found on github.
Authors: Tainyi Zhang, Zheng-Peng Duan, Peng-Tao Jiang, Bo Li, Ming-Ming Cheng, Chun-Le Guo, Chongyi Li
Abstract: Diffusion-based real-world image super-resolution (Real-ISR) methods have demonstrated impressive performance. To achieve efficient Real-ISR, many works employ Variational Score Distillation (VSD) to distill pre-trained stable-diffusion (SD) model for one-step SR with a fixed timestep. However, due to the different noise injection timesteps, the SD will perform different generative priors. Therefore, a fixed timestep is difficult for these methods to fully leverage the generative priors in SD, leading to suboptimal performance. To address this, we propose a Time-Aware one-step Diffusion Network for Real-ISR (TADSR). We first introduce a Time-Aware VAE Encoder, which projects the same image into different latent features based on timesteps. Through joint dynamic variation of timesteps and latent features, the student model can better align with the input pattern distribution of the pre-trained SD, thereby enabling more effective utilization of SD's generative capabilities. To better activate the generative prior of SD at different timesteps, we propose a Time-Aware VSD loss that bridges the timesteps of the student model and those of the teacher model, thereby producing more consistent generative prior guidance conditioned on timesteps. Additionally, though utilizing the generative prior in SD at different timesteps, our method can naturally achieve controllable trade-offs between fidelity and realism by changing the timestep condition. Experimental results demonstrate that our method achieves both state-of-the-art performance and controllable SR results with only a single step.