Peer Revenue Returns subject to peer review

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Computer Vision

cs.CV Watch Computer Vision

Friday 16 January

Computer Vision

IDDR-NGP: Incorporating Detectors for Distractor Removal with Instant Neural Radiance Field

Xianliang Huang, Jiajie Gou, Shuhang Chen et al. (+3 more)

This paper presents the first unified distractor removal method, named IDDR-NGP, which directly operates on Instant-NPG. The method is able to remove a wide range of distractors in 3D scenes, such as snowflakes, confetti, defoliation and petals, whereas existing methods usually focus on a specific type of distractors. By incorporating implicit 3D representations with 2D detectors, we demonstrate that it is possible to efficiently restore 3D scenes from multiple corrupted images. We design the learned perceptual image patch similarity~( LPIPS) loss and the multi-view compensation loss (MVCL) to jointly optimize the rendering results of IDDR-NGP, which could aggregate information from multi-view corrupted images. All of them can be trained in an end-to-end manner to synthesize high-quality 3D scenes. To support the research on distractors removal in implicit 3D representations, we build a new benchmark dataset that consists of both synthetic and real-world distractors. To validate the effectiveness and robustness of IDDR-NGP, we provide a wide range of distractors with corresponding annotated labels added to both realistic and synthetic scenes. Extensive experimental results demonstrate the effectiveness and robustness of IDDR-NGP in removing multiple types of distractors. In addition, our approach achieves results comparable with the existing SOTA desnow methods and is capable of accurately removing both realistic and synthetic distractors.
Read the full abstract

All authors Xianliang Huang, Jiajie Gou, Shuhang Chen, Zhizhou Zhong, Jihong Guan, Shuigeng Zhou

Computer Vision · Cross-listed from cs.LG

Matching High-Dimensional Geometric Quantiles for Test-Time Adaptation of Transformers and Convolutional Networks Alike

Sravan Danda, Aditya Challa, Shlok Mehendale et al. (+1 more)

Test-time adaptation (TTA) refers to adapting a classifier for the test data when the probability distribution of the test data slightly differs from that of the training data of the model. To the best of our knowledge, most of the existing TTA approaches modify the weights of the classifier relying heavily on the architecture. It is unclear as to how these approaches are extendable to generic architectures. In this article, we propose an architecture-agnostic approach to TTA by adding an adapter network pre-processing the input images suitable to the classifier. This adapter is trained using the proposed quantile loss. Unlike existing approaches, we correct for the distribution shift by matching high-dimensional geometric quantiles. We prove theoretically that under suitable conditions minimizing quantile loss can learn the optimal adapter. We validate our approach on CIFAR10-C, CIFAR100-C and TinyImageNet-C by training both classic convolutional and transformer networks on CIFAR10, CIFAR100 and TinyImageNet datasets.
Read the full abstract

All authors Sravan Danda, Aditya Challa, Shlok Mehendale, Snehanshu Saha

Computer Vision · Cross-listed from q-bio.NC

KOCOBrain: Kuramoto-Guided Graph Network for Uncovering Structure-Function Coupling in Adolescent Prenatal Drug Exposure

Badhan Mazumder, Lei Wu, Sir-Lord Wiafe et al. (+2 more)

Exposure to psychoactive substances during pregnancy, such as cannabis, can disrupt neurodevelopment and alter large-scale brain networks, yet identifying their neural signatures remains challenging. We introduced KOCOBrain: KuramotO COupled Brain Graph Network; a unified graph neural network framework that integrates structural and functional connectomes via Kuramoto-based phase dynamics and cognition-aware attention. The Kuramoto layer models neural synchronization over anatomical connections, generating phase-informed embeddings that capture structure-function coupling, while cognitive scores modulate information routing in a subject-specific manner followed by a joint objective enhancing robustness under class imbalance scenario. Applied to the ABCD cohort, KOCOBrain improved prenatal drug exposure prediction over relevant baselines and revealed interpretable structure-function patterns that reflect disrupted brain network coordination associated with early exposure.
Read the full abstract

All authors Badhan Mazumder, Lei Wu, Sir-Lord Wiafe, Vince D. Calhoun, Dong Hye Ye

Computer Vision

MMedExpert-R1: Strengthening Multimodal Medical Reasoning via Domain-Specific Adaptation and Clinical Guideline Reinforcement

Meidan Ding, Jipeng Zhang, Wenxuan Wang et al. (+4 more)

Medical Vision-Language Models (MedVLMs) excel at perception tasks but struggle with complex clinical reasoning required in real-world scenarios. While reinforcement learning (RL) has been explored to enhance reasoning capabilities, existing approaches face critical mismatches: the scarcity of deep reasoning data, cold-start limits multi-specialty alignment, and standard RL algorithms fail to model clinical reasoning diversity. We propose MMedExpert-R1, a novel reasoning MedVLM that addresses these challenges through domain-specific adaptation and clinical guideline reinforcement. We construct MMedExpert, a high-quality dataset of 10K samples across four specialties with step-by-step reasoning traces. Our Domain-Specific Adaptation (DSA) creates specialty-specific LoRA modules to provide diverse initialization, while Guideline-Based Advantages (GBA) explicitly models different clinical reasoning perspectives to align with real-world diagnostic strategies. Conflict-Aware Capability Integration then merges these specialized experts into a unified agent, ensuring robust multi-specialty alignment. Comprehensive experiments demonstrate state-of-the-art performance, with our 7B model achieving 27.50 on MedXpert-MM and 83.03 on OmniMedVQA, establishing a robust foundation for reliable multimodal medical reasoning systems.
Read the full abstract

All authors Meidan Ding, Jipeng Zhang, Wenxuan Wang, Haiqin Zhong, Xiaoling Luo, Wenting Chen, Linlin Shen

Computer Vision

PatientVLM Meets DocVLM: Pre-Consultation Dialogue Between Vision-Language Models for Efficient Diagnosis

K Lokesh, Abhirama Subramanyam Penamakuri, Uday Agarwal et al. (+4 more)

Traditionally, AI research in medical diagnosis has largely centered on image analysis. While this has led to notable advancements, the absence of patient-reported symptoms continues to hinder diagnostic accuracy. To address this, we propose a Pre-Consultation Dialogue Framework (PCDF) that mimics real-world diagnostic procedures, where doctors iteratively query patients before reaching a conclusion. Specifically, we simulate diagnostic dialogues between two vision-language models (VLMs): a DocVLM, which generates follow-up questions based on the image and dialogue history, and a PatientVLM, which responds using a symptom profile derived from the ground-truth diagnosis. We additionally conducted a small-scale clinical validation of the synthetic symptoms generated by our framework, with licensed clinicians confirming their clinical relevance, symptom coverage, and overall realism. These findings indicate that the resulting DocVLM-PatientVLM interactions form coherent, multi-turn consultations paired with images and diagnoses, which we then use to fine-tune the DocVLM. This dialogue-based supervision leads to substantial gains over image-only training, highlighting the value of realistic symptom elicitation for diagnosis.
Read the full abstract

All authors K Lokesh, Abhirama Subramanyam Penamakuri, Uday Agarwal, Apoorva Challa, Shreya K Gowda, Somesh Gupta, Anand Mishra

Computer Vision

Sparse Data Tree Canopy Segmentation: Fine-Tuning Leading Pretrained Models on Only 150 Images

David Szczecina, Hudson Sun, Anthony Bertnyk et al. (+3 more)

Tree canopy detection from aerial imagery is an important task for environmental monitoring, urban planning, and ecosystem analysis. Simulating real-life data annotation scarcity, the Solafune Tree Canopy Detection competition provides a small and imbalanced dataset of only 150 annotated images, posing significant challenges for training deep models without severe overfitting. In this work, we evaluate five representative architectures, YOLOv11, Mask R-CNN, DeepLabv3, Swin-UNet, and DINOv2, to assess their suitability for canopy segmentation under extreme data scarcity. Our experiments show that pretrained convolution-based models, particularly YOLOv11 and Mask R-CNN, generalize significantly better than pretrained transformer-based models. DeeplabV3, Swin-UNet and DINOv2 underperform likely due to differences between semantic and instance segmentation tasks, the high data requirements of Vision Transformers, and the lack of strong inductive biases. These findings confirm that transformer-based architectures struggle in low-data regimes without substantial pretraining or augmentation and that differences between semantic and instance segmentation further affect model performance. We provide a detailed analysis of training strategies, augmentation policies, and model behavior under the small-data constraint and demonstrate that lightweight CNN-based methods remain the most reliable for canopy detection on limited imagery.
Read the full abstract

All authors David Szczecina, Hudson Sun, Anthony Bertnyk, Niloofar Azad, Kyle Gao, Lincoln Linlin Xu

Computer Vision

RobuMTL: Enhancing Multi-Task Learning Robustness Against Weather Conditions

Tasneem Shaffee and Sherief Reda

Robust Multi-Task Learning (MTL) is crucial for autonomous systems operating in real-world environments, where adverse weather conditions can severely degrade model performance and reliability. In this paper, we introduce RobuMTL, a novel architecture designed to adaptively address visual degradation by dynamically selecting task-specific hierarchical Low-Rank Adaptation (LoRA) modules and a LoRA expert squad based on input perturbations in a mixture-of-experts fashion. Our framework enables adaptive specialization based on input characteristics, improving robustness across diverse real-world conditions. To validate our approach, we evaluated it on the PASCAL and NYUD-v2 datasets and compared it against single-task models, standard MTL baselines, and state-of-the-art methods. On the PASCAL benchmark, RobuMTL delivers a +2.8% average relative improvement under single perturbations and up to +44.4% under mixed weather conditions compared to the MTL baseline. On NYUD-v2, RobuMTL achieves a +9.7% average relative improvement across tasks. The code is available at GitHub.
Read the full abstract

Computer Vision

Self-learned representation-guided latent diffusion model for breast cancer classification in deep ultraviolet whole surface images

Pouya Afshin, David Helminiak, Tianling Niu et al. (+4 more)

Breast-Conserving Surgery (BCS) requires precise intraoperative margin assessment to preserve healthy tissue. Deep Ultraviolet Fluorescence Scanning Microscopy (DUV-FSM) offers rapid, high-resolution surface imaging for this purpose; however, the scarcity of annotated DUV data hinders the training of robust deep learning models. To address this, we propose an Self-Supervised Learning (SSL)-guided Latent Diffusion Model (LDM) to generate high-quality synthetic training patches. By guiding the LDM with embeddings from a fine-tuned DINO teacher, we inject rich semantic details of cellular structures into the synthetic data. We combine real and synthetic patches to fine-tune a Vision Transformer (ViT), utilizing patch prediction aggregation for WSI-level classification. Experiments using 5-fold cross-validation demonstrate that our method achieves 96.47 % accuracy and reduces the FID score to 45.72, significantly outperforming class-conditioned baselines.
Read the full abstract

All authors Pouya Afshin, David Helminiak, Tianling Niu, Julie M. Jorns, Tina Yen, Bing Yu, Dong Hye Ye

Computer Vision

Classification of Chest XRay Diseases through image processing and analysis techniques

Santiago Martínez Novoa, María Catalina Ibáñez, Lina Gómez Mesa et al. (+1 more)

Multi-Classification Chest X-Ray Images are one of the most prevalent forms of radiological examination used for diagnosing thoracic diseases. In this study, we offer a concise overview of several methods employed for tackling this task, including DenseNet121. In addition, we deploy an open-source web-based application. In our study, we conduct tests to compare different methods and see how well they work. We also look closely at the weaknesses of the methods we propose and suggest ideas for making them better in the future. Our code is available at: https://github.com/AML4206-MINE20242/Proyecto_AML
Read the full abstract

All authors Santiago Martínez Novoa, María Catalina Ibáñez, Lina Gómez Mesa, Jeremias Kramer

Thursday 15 January

Computer Vision · Cross-listed from cs.RO

FeudalNav: A Simple Framework for Visual Navigation

Faith Johnson, Bryan Bo Cao, Shubham Jain et al. (+2 more)

Visual navigation for robotics is inspired by the human ability to navigate environments using visual cues and memory, eliminating the need for detailed maps. In unseen, unmapped, or GPS-denied settings, traditional metric map-based methods fall short, prompting a shift toward learning-based approaches with minimal exploration. In this work, we develop a hierarchical framework that decomposes the navigation decision-making process into multiple levels. Our method learns to select subgoals through a simple, transferable waypoint selection network. A key component of the approach is a latent-space memory module organized solely by visual similarity, as a proxy for distance. This alternative to graph-based topological representations proves sufficient for navigation tasks, providing a compact, light-weight, simple-to-train navigator that can find its way to the goal in novel locations. We show competitive results with a suite of SOTA methods in Habitat AI environments without using any odometry in training or inference. An additional contribution leverages the interpretablility of the framework for interactive navigation. We consider the question: how much direction intervention/interaction is needed to achieve success in all trials? We demonstrate that even minimal human involvement can significantly enhance overall navigation performance.
Read the full abstract

All authors Faith Johnson, Bryan Bo Cao, Shubham Jain, Ashwin Ashok, Kristin Dana

Computer Vision

Human detectors are surprisingly powerful reward models

Kumar Ashutosh, XuDong Wang, Xi Yin et al. (+4 more)

Video generation models have recently achieved impressive visual fidelity and temporal coherence. Yet, they continue to struggle with complex, non-rigid motions, especially when synthesizing humans performing dynamic actions such as sports, dance, etc. Generated videos often exhibit missing or extra limbs, distorted poses, or physically implausible actions. In this work, we propose a remarkably simple reward model, HuDA, to quantify and improve the human motion in generated videos. HuDA integrates human detection confidence for appearance quality, and a temporal prompt alignment score to capture motion realism. We show this simple reward function that leverages off-the-shelf models without any additional training, outperforms specialized models finetuned with manually annotated data. Using HuDA for Group Reward Policy Optimization (GRPO) post-training of video models, we significantly enhance video generation, especially when generating complex human motions, outperforming state-of-the-art models like Wan 2.1, with win-rate of 73%. Finally, we demonstrate that HuDA improves generation quality beyond just humans, for instance, significantly improving generation of animal videos and human-object interactions.
Read the full abstract

All authors Kumar Ashutosh, XuDong Wang, Xi Yin, Kristen Grauman, Adam Polyak, Ishan Misra, Rohit Girdhar

Computer Vision

LTV-YOLO: A Lightweight Thermal Object Detector for Young Pedestrians in Adverse Conditions

Abdullah Jirjees, Ryan Myers, Muhammad Haris Ikram et al. (+1 more)

Detecting vulnerable road users (VRUs), particularly children and adolescents, in low light and adverse weather conditions remains a critical challenge in computer vision, surveillance, and autonomous vehicle systems. This paper presents a purpose-built lightweight object detection model designed to identify young pedestrians in various environmental scenarios. To address these challenges, our approach leverages thermal imaging from long-wave infrared (LWIR) cameras, which enhances detection reliability in conditions where traditional RGB cameras operating in the visible spectrum fail. Based on the YOLO11 architecture and customized for thermal detection, our model, termed LTV-YOLO (Lightweight Thermal Vision YOLO), is optimized for computational efficiency, accuracy and real-time performance on edge devices. By integrating separable convolutions in depth and a feature pyramid network (FPN), LTV-YOLO achieves strong performance in detecting small-scale, partially occluded, and thermally distinct VRUs while maintaining a compact architecture. This work contributes a practical and scalable solution to improve pedestrian safety in intelligent transportation systems, particularly in school zones, autonomous navigation, and smart city infrastructure. Unlike prior thermal detectors, our contribution is task-specific: a thermally only edge-capable design designed for young and small VRUs (children and distant adults). Although FPN and depthwise separable convolutions are standard components, their integration into a thermal-only pipeline optimized for short/occluded VRUs under adverse conditions is, to the best of our knowledge, novel.
Read the full abstract

All authors Abdullah Jirjees, Ryan Myers, Muhammad Haris Ikram, Mohamed H. Zaki

Computer Vision

Zeros can be Informative: Masked Binary U-Net for Image Segmentation on Tensor Cores

Chunshu Wu, Ruibing Song, Sushant Kondguli et al. (+2 more)

Real-time image segmentation is a key enabler for AR/VR, robotics, drones, and autonomous systems, where tight accuracy, latency, and energy budgets must be met on resource-constrained edge devices. While U-Net offers a favorable balance of accuracy and efficiency compared to large transformer-based models, achieving real-time performance on high-resolution input remains challenging due to compute, memory, and power limits. Extreme quantization, particularly binary networks, is appealing for its hardware-friendly operations. However, two obstacles limit practicality: (1) severe accuracy degradation, and (2) a lack of end-to-end implementations that deliver efficiency on general-purpose GPUs. We make two empirical observations that guide our design. (1) An explicit zero state is essential: training with zero masking to binary U-Net weights yields noticeable sparsity. (2) Quantization sensitivity is uniform across layers. Motivated by these findings, we introduce Masked Binary U-Net (MBU-Net), obtained through a cost-aware masking strategy that prioritizes masking where it yields the highest accuracy-per-cost, reconciling accuracy with near-binary efficiency. To realize these gains in practice, we develop a GPU execution framework that maps MBU-Net to Tensor Cores via a subtractive bit-encoding scheme, efficiently implementing masked binary weights with binary activations. This design leverages native binary Tensor Core BMMA instructions, enabling high throughput and energy savings on widely available GPUs. Across 3 segmentation benchmarks, MBU-Net attains near full-precision accuracy (3% average drop) while delivering 2.04x speedup and 3.54x energy reductions over a 16-bit floating point U-Net.
Read the full abstract

All authors Chunshu Wu, Ruibing Song, Sushant Kondguli, Tong Geng, Ang Li

Computer Vision · Cross-listed from cs.NI

An Efficient and Explainable KAN Framework for Wireless Radiation Field Prediction

Jingzhou Shen and Xuyu Wang

Modeling wireless channels accurately remains a challenge due to environmental variations and signal uncertainties. Recent neural networks can learn radio frequency~(RF) signal propagation patterns, but they process each voxel on the ray independently, without considering global context or environmental factors. Our paper presents a new approach that learns comprehensive representations of complete rays rather than individual points, capturing more detailed environmental features. We integrate a Kolmogorov-Arnold network (KAN) architecture with transformer modules to achieve better performance across realistic and synthetic scenes while maintaining computational efficiency. Our experimental results show that this approach outperforms existing methods in various scenarios. Ablation studies confirm that each component of our model contributes to its effectiveness. Additional experiments provide clear explanations for our model's performance.
Read the full abstract

Computer Vision

PSSI-MaxST: An Efficient Pixel-Segment Similarity Index Using Intensity and Smoothness Features for Maximum Spanning Tree Based Segmentation

Kaustubh Shivshankar Shejole and Gaurav Mishra

Interactive graph-based segmentation methods partition an image into foreground and background regions with the aid of user inputs. However, existing approaches often suffer from high computational costs, sensitivity to user interactions, and degraded performance when the foreground and background share similar color distributions. A key factor influencing segmentation performance is the similarity measure used for assigning edge weights in the graph. To address these challenges, we propose a novel Pixel Segment Similarity Index (PSSI), which leverages the harmonic mean of inter-channel similarities by incorporating both pixel intensity and spatial smoothness features. The harmonic mean effectively penalizes dissimilarities in any individual channel, enhancing robustness. The computational complexity of PSSI is $\mathcal{O}(B)$, where $B$ denotes the number of histogram bins. Our segmentation framework begins with low-level segmentation using MeanShift, which effectively captures color, texture, and segment shape. Based on the resulting pixel segments, we construct a pixel-segment graph with edge weights determined by PSSI. For partitioning, we employ the Maximum Spanning Tree (MaxST), which captures strongly connected local neighborhoods beneficial for precise segmentation. The integration of the proposed PSSI, MeanShift, and MaxST allows our method to jointly capture color similarity, smoothness, texture, shape, and strong local connectivity. Experimental evaluations on the GrabCut and Images250 datasets demonstrate that our method consistently outperforms current graph-based interactive segmentation methods such as AMOE, OneCut, and SSNCut in terms of segmentation quality, as measured by Jaccard Index (IoU), $F_1$ score, execution time and Mean Error (ME). Code is publicly available at: https://github.com/KaustubhShejole/PSSI-MaxST.
Read the full abstract

Computer Vision

Aesthetics as Structural Harm: Algorithmic Lookism Across Text-to-Image Generation and Classification

Miriam Doh, Aditya Gulati, Corinna Canali et al. (+1 more)

This paper examines algorithmic lookism-the systematic preferential treatment based on physical appearance-in text-to-image (T2I) generative AI and a downstream gender classification task. Through the analysis of 26,400 synthetic faces created with Stable Diffusion 2.1 and 3.5 Medium, we demonstrate how generative AI models systematically associate facial attractiveness with positive attributes and vice-versa, mirroring socially constructed biases rather than evidence-based correlations. Furthermore, we find significant gender bias in three gender classification algorithms depending on the attributes of the input faces. Our findings reveal three critical harms: (1) the systematic encoding of attractiveness-positive attribute associations in T2I models; (2) gender disparities in classification systems, where women's faces, particularly those generated with negative attributes, suffer substantially higher misclassification rates than men's; and (3) intensifying aesthetic constraints in newer models through age homogenization, gendered exposure patterns, and geographic reductionism. These convergent patterns reveal algorithmic lookism as systematic infrastructure operating across AI vision systems, compounding existing inequalities through both representation and recognition. Disclaimer: This work includes visual and textual content that reflects stereotypical associations between physical appearance and socially constructed attributes, including gender, race, and traits associated with social desirability. Any such associations found in this study emerge from the biases embedded in generative AI systems-not from empirical truths or the authors' views.
Read the full abstract

All authors Miriam Doh, Aditya Gulati, Corinna Canali, Nuria Oliver

Computer Vision

FrankenMotion: Part-level Human Motion Generation and Composition

Chuqiao Li, Xianghui Xie, Yong Cao et al. (+2 more)

Human motion generation from text prompts has made remarkable progress in recent years. However, existing methods primarily rely on either sequence-level or action-level descriptions due to the absence of fine-grained, part-level motion annotations. This limits their controllability over individual body parts. In this work, we construct a high-quality motion dataset with atomic, temporally-aware part-level text annotations, leveraging the reasoning capabilities of large language models (LLMs). Unlike prior datasets that either provide synchronized part captions with fixed time segments or rely solely on global sequence labels, our dataset captures asynchronous and semantically distinct part movements at fine temporal resolution. Based on this dataset, we introduce a diffusion-based part-aware motion generation framework, namely FrankenMotion, where each body part is guided by its own temporally-structured textual prompt. This is, to our knowledge, the first work to provide atomic, temporally-aware part-level motion annotations and have a model that allows motion generation with both spatial (body part) and temporal (atomic action) control. Experiments demonstrate that FrankenMotion outperforms all previous baseline models adapted and retrained for our setting, and our model can compose motions unseen during training. Our code and dataset will be publicly available upon publication.
Read the full abstract

All authors Chuqiao Li, Xianghui Xie, Yong Cao, Andreas Geiger, Gerard Pons-Moll

Computer Vision

Medical SAM3: A Foundation Model for Universal Prompt-Driven Medical Image Segmentation

Chongcong Jiang, Tianxingjian Ding, Chuhan Song et al. (+7 more)

Promptable segmentation foundation models such as SAM3 have demonstrated strong generalization capabilities through interactive and concept-based prompting. However, their direct applicability to medical image segmentation remains limited by severe domain shifts, the absence of privileged spatial prompts, and the need to reason over complex anatomical and volumetric structures. Here we present Medical SAM3, a foundation model for universal prompt-driven medical image segmentation, obtained by fully fine-tuning SAM3 on large-scale, heterogeneous 2D and 3D medical imaging datasets with paired segmentation masks and text prompts. Through a systematic analysis of vanilla SAM3, we observe that its performance degrades substantially on medical data, with its apparent competitiveness largely relying on strong geometric priors such as ground-truth-derived bounding boxes. These findings motivate full model adaptation beyond prompt engineering alone. By fine-tuning SAM3's model parameters on 33 datasets spanning 10 medical imaging modalities, Medical SAM3 acquires robust domain-specific representations while preserving prompt-driven flexibility. Extensive experiments across organs, imaging modalities, and dimensionalities demonstrate consistent and significant performance gains, particularly in challenging scenarios characterized by semantic ambiguity, complex morphology, and long-range 3D context. Our results establish Medical SAM3 as a universal, text-guided segmentation foundation model for medical imaging and highlight the importance of holistic model adaptation for achieving robust prompt-driven segmentation under severe domain shift. Code and model will be made available at https://github.com/AIM-Research-Lab/Medical-SAM3.
Read the full abstract

All authors Chongcong Jiang, Tianxingjian Ding, Chuhan Song, Jiachen Tu, Ziyang Yan, Yihua Shao, Zhenyi Wang, Yuzhang Shang, Tianyu Han, Yu Tian

Computer Vision

Effects of Different Attention Mechanisms Applied on 3D Models in Video Classification

Mohammad Rasras, Iuliana Marin, Serban Radu et al. (+1 more)

Human action recognition has become an important research focus in computer vision due to the wide range of applications where it is used. 3D Resnet-based CNN models, particularly MC3, R3D, and R(2+1)D, have different convolutional filters to extract spatiotemporal features. This paper investigates the impact of reducing the captured knowledge from temporal data, while increasing the resolution of the frames. To establish this experiment, we created similar designs to the three originals, but with a dropout layer added before the final classifier. Secondly, we then developed ten new versions for each one of these three designs. The variants include special attention blocks within their architecture, such as convolutional block attention module (CBAM), temporal convolution networks (TCN), in addition to multi-headed and channel attention mechanisms. The purpose behind that is to observe the extent of the influence each of these blocks has on performance for the restricted-temporal models. The results of testing all the models on UCF101 have shown accuracy of 88.98% for the variant with multiheaded attention added to the modified R(2+1)D. This paper concludes the significance of missing temporal features in the performance of the newly created increased resolution models. The variants had different behavior on class-level accuracy, despite the similarity of their enhancements to the overall performance.
Read the full abstract

All authors Mohammad Rasras, Iuliana Marin, Serban Radu, Irina Mocanu

Computer Vision

One Model, Many Behaviors: Training-Induced Effects on Out-of-Distribution Detection

Gerhard Krumpl, Henning Avenhaus and Horst Possegger

Out-of-distribution (OOD) detection is crucial for deploying robust and reliable machine-learning systems in open-world settings. Despite steady advances in OOD detectors, their interplay with modern training pipelines that maximize in-distribution (ID) accuracy and generalization remains under-explored. We investigate this link through a comprehensive empirical study. Fixing the architecture to the widely adopted ResNet-50, we benchmark 21 post-hoc, state-of-the-art OOD detection methods across 56 ImageNet-trained models obtained via diverse training strategies and evaluate them on eight OOD test sets. Contrary to the common assumption that higher ID accuracy implies better OOD detection performance, we uncover a non-monotonic relationship: OOD performance initially improves with accuracy but declines once advanced training recipes push accuracy beyond the baseline. Moreover, we observe a strong interdependence between training strategy, detector choice, and resulting OOD performance, indicating that no single method is universally optimal.
Read the full abstract

Computer Vision

Can Vision-Language Models Understand Construction Workers? An Exploratory Study

Hieu Bui, Nathaniel E. Chodosh and Arash Tavakoli

As robotics become increasingly integrated into construction workflows, their ability to interpret and respond to human behavior will be essential for enabling safe and effective collaboration. Vision-Language Models (VLMs) have emerged as a promising tool for visual understanding tasks and offer the potential to recognize human behaviors without extensive domain-specific training. This capability makes them particularly appealing in the construction domain, where labeled data is scarce and monitoring worker actions and emotional states is critical for safety and productivity. In this study, we evaluate the performance of three leading VLMs, GPT-4o, Florence 2, and LLaVa-1.5, in detecting construction worker actions and emotions from static site images. Using a curated dataset of 1,000 images annotated across ten action and ten emotion categories, we assess each model's outputs through standardized inference pipelines and multiple evaluation metrics. GPT-4o consistently achieved the highest scores across both tasks, with an average F1-score of 0.756 and accuracy of 0.799 in action recognition, and an F1-score of 0.712 and accuracy of 0.773 in emotion recognition. Florence 2 performed moderately, with F1-scores of 0.497 for action and 0.414 for emotion, while LLaVa-1.5 showed the lowest overall performance, with F1-scores of 0.466 for action and 0.461 for emotion. Confusion matrix analyses revealed that all models struggled to distinguish semantically close categories, such as collaborating in teams versus communicating with supervisors. While the results indicate that general-purpose VLMs can offer a baseline capability for human behavior recognition in construction environments, further improvements, such as domain adaptation, temporal modeling, or multimodal sensing, may be needed for real-world reliability.
Read the full abstract

Computer Vision

A Unified 3D Object Perception Framework for Real-Time Outside-In Multi-Camera Systems

Yizhou Wang, Sameer Pusegaonkar, Yuxing Wang et al. (+10 more)

Accurate 3D object perception and multi-target multi-camera (MTMC) tracking are fundamental for the digital transformation of industrial infrastructure. However, transitioning "inside-out" autonomous driving models to "outside-in" static camera networks presents significant challenges due to heterogeneous camera placements and extreme occlusion. In this paper, we present an adapted Sparse4D framework specifically optimized for large-scale infrastructure environments. Our system leverages absolute world-coordinate geometric priors and introduces an occlusion-aware ReID embedding module to maintain identity stability across distributed sensor networks. To bridge the Sim2Real domain gap without manual labeling, we employ a generative data augmentation strategy using the NVIDIA COSMOS framework, creating diverse environmental styles that enhance the model's appearance-invariance. Evaluated on the AI City Challenge 2025 benchmark, our camera-only framework achieves a state-of-the-art HOTA of $45.22$. Furthermore, we address real-time deployment constraints by developing an optimized TensorRT plugin for Multi-Scale Deformable Aggregation (MSDA). Our hardware-accelerated implementation achieves a $2.15\times$ speedup on modern GPU architectures, enabling a single Blackwell-class GPU to support over 64 concurrent camera streams.
Read the full abstract

All authors Yizhou Wang, Sameer Pusegaonkar, Yuxing Wang, Anqi Li, Vishal Kumar, Chetan Sethi, Ganapathy Aiyer, Yun He, Kartikay Thakkar, Swapnil Rathi, Bhushan Rupde, Zheng Tang, Sujit Biswas

Computer Vision

ICONIC-444: A 3.1-Million-Image Dataset for OOD Detection Research

Gerhard Krumpl, Henning Avenhaus and Horst Possegger

Current progress in out-of-distribution (OOD) detection is limited by the lack of large, high-quality datasets with clearly defined OOD categories across varying difficulty levels (near- to far-OOD) that support both fine- and coarse-grained computer vision tasks. To address this limitation, we introduce ICONIC-444 (Image Classification and OOD Detection with Numerous Intricate Complexities), a specialized large-scale industrial image dataset containing over 3.1 million RGB images spanning 444 classes tailored for OOD detection research. Captured with a prototype industrial sorting machine, ICONIC-444 closely mimics real-world tasks. It complements existing datasets by offering structured, diverse data suited for rigorous OOD evaluation across a spectrum of task complexities. We define four reference tasks within ICONIC-444 to benchmark and advance OOD detection research and provide baseline results for 22 state-of-the-art post-hoc OOD detection methods.
Read the full abstract

Computer Vision

Future Optical Flow Prediction Improves Robot Control & Video Generation

Kanchana Ranasinghe, Honglu Zhou, Yu Fang et al. (+7 more)

Future motion representations, such as optical flow, offer immense value for control and generative tasks. However, forecasting generalizable spatially dense motion representations remains a key challenge, and learning such forecasting from noisy, real-world data remains relatively unexplored. We introduce FOFPred, a novel language-conditioned optical flow forecasting model featuring a unified Vision-Language Model (VLM) and Diffusion architecture. This unique combination enables strong multimodal reasoning with pixel-level generative fidelity for future motion prediction. Our model is trained on web-scale human activity data-a highly scalable but unstructured source. To extract meaningful signals from this noisy video-caption data, we employ crucial data preprocessing techniques and our unified architecture with strong image pretraining. The resulting trained model is then extended to tackle two distinct downstream tasks in control and generation. Evaluations across robotic manipulation and video generation under language-driven settings establish the cross-domain versatility of FOFPred, confirming the value of a unified VLM-Diffusion architecture and scalable learning from diverse web data for future optical flow prediction.
Read the full abstract

All authors Kanchana Ranasinghe, Honglu Zhou, Yu Fang, Luyu Yang, Le Xue, Ran Xu, Caiming Xiong, Silvio Savarese, Michael S Ryoo, Juan Carlos Niebles

Computer Vision

WildRayZer: Self-supervised Large View Synthesis in Dynamic Environments

Xuweiyi Chen, Wentao Zhou and Zezhou Cheng

We present WildRayZer, a self-supervised framework for novel view synthesis (NVS) in dynamic environments where both the camera and objects move. Dynamic content breaks the multi-view consistency that static NVS models rely on, leading to ghosting, hallucinated geometry, and unstable pose estimation. WildRayZer addresses this by performing an analysis-by-synthesis test: a camera-only static renderer explains rigid structure, and its residuals reveal transient regions. From these residuals, we construct pseudo motion masks, distill a motion estimator, and use it to mask input tokens and gate loss gradients so supervision focuses on cross-view background completion. To enable large-scale training and evaluation, we curate Dynamic RealEstate10K (D-RE10K), a real-world dataset of 15K casually captured dynamic sequences, and D-RE10K-iPhone, a paired transient and clean benchmark for sparse-view transient-aware NVS. Experiments show that WildRayZer consistently outperforms optimization-based and feed-forward baselines in both transient-region removal and full-frame NVS quality with a single feed-forward pass.
Read the full abstract

Computer Vision

Alterbute: Editing Intrinsic Attributes of Objects in Images

Tal Reiss, Daniel Winter, Matan Cohen et al. (+4 more)

We introduce Alterbute, a diffusion-based method for editing an object's intrinsic attributes in an image. We allow changing color, texture, material, and even the shape of an object, while preserving its perceived identity and scene context. Existing approaches either rely on unsupervised priors that often fail to preserve identity or use overly restrictive supervision that prevents meaningful intrinsic variations. Our method relies on: (i) a relaxed training objective that allows the model to change both intrinsic and extrinsic attributes conditioned on an identity reference image, a textual prompt describing the target intrinsic attributes, and a background image and object mask defining the extrinsic context. At inference, we restrict extrinsic changes by reusing the original background and object mask, thereby ensuring that only the desired intrinsic attributes are altered; (ii) Visual Named Entities (VNEs) - fine-grained visual identity categories (e.g., ”Porsche 911 Carrera”) that group objects sharing identity-defining features while allowing variation in intrinsic attributes. We use a vision-language model to automatically extract VNE labels and intrinsic attribute descriptions from a large public image dataset, enabling scalable, identity-preserving supervision. Alterbute outperforms existing methods on identity-preserving object intrinsic attribute editing.
Read the full abstract

All authors Tal Reiss, Daniel Winter, Matan Cohen, Alex Rav-Acha, Yael Pritch, Ariel Shamir, Yedid Hoshen

Computer Vision

From One-to-One to Many-to-Many: Dynamic Cross-Layer Injection for Deep Vision-Language Fusion

Cheng Chen, Yuyu Guo, Pengpeng Zeng et al. (+4 more)

Vision-Language Models (VLMs) create a severe visual feature bottleneck by using a crude, asymmetric connection that links only the output of the vision encoder to the input of the large language model (LLM). This static architecture fundamentally limits the ability of LLMs to achieve comprehensive alignment with hierarchical visual knowledge, compromising their capacity to accurately integrate local details with global semantics into coherent reasoning. To resolve this, we introduce Cross-Layer Injection (CLI), a novel and lightweight framework that forges a dynamic many-to-many bridge between the two modalities. CLI consists of two synergistic, parameter-efficient components: an Adaptive Multi-Projection (AMP) module that harmonizes features from diverse vision layers, and an Adaptive Gating Fusion (AGF) mechanism that empowers the LLM to selectively inject the most relevant visual information based on its real-time decoding context. We validate the effectiveness and versatility of CLI by integrating it into LLaVA-OneVision and LLaVA-1.5. Extensive experiments on 18 diverse benchmarks demonstrate significant performance improvements, establishing CLI as a scalable paradigm that unlocks deeper multimodal understanding by granting LLMs on-demand access to the full visual hierarchy.
Read the full abstract

All authors Cheng Chen, Yuyu Guo, Pengpeng Zeng, Jingkuan Song, Peng Di, Hang Yu, Lianli Gao

Computer Vision

See Less, Drive Better: Generalizable End-to-End Autonomous Driving via Foundation Models Stochastic Patch Selection

Amir Mallak, Erfan Aasi, Shiva Sreeram et al. (+3 more)

Recent advances in end-to-end autonomous driving show that policies trained on patch-aligned features extracted from foundation models generalize better to Out-of-Distribution (OOD). We hypothesize that due to the self-attention mechanism, each patch feature implicitly embeds/contains information from all other patches, represented in a different way and intensity, making these descriptors highly redundant. We quantify redundancy in such (BLIP2) features via PCA and cross-patch similarity: $90$% of variance is captured by $17/64$ principal components, and strong inter-token correlations are pervasive. Training on such overlapping information leads the policy to overfit spurious correlations, hurting OOD robustness. We present Stochastic-Patch-Selection (SPS), a simple yet effective approach for learning policies that are more robust, generalizable, and efficient. For every frame, SPS randomly masks a fraction of patch descriptors, not feeding them to the policy model, while preserving the spatial layout of the remaining patches. Thus, the policy is provided with different stochastic but complete views of the (same) scene: every random subset of patches acts like a different, yet still sensible, coherent projection of the world. The policy thus bases its decisions on features that are invariant to which specific tokens survive. Extensive experiments confirm that across all OOD scenarios, our method outperforms the state of the art (SOTA), achieving a $6.2$% average improvement and up to $20.4$% in closed-loop simulations, while being $2.4\times$ faster. We conduct ablations over masking rates and patch-feature reorganization, training and evaluating 9 systems, with 8 of them surpassing prior SOTA. Finally, we show that the same learned policy transfers to a physical, real-world car without any tuning.
Read the full abstract

All authors Amir Mallak, Erfan Aasi, Shiva Sreeram, Tsun-Hsuan Wang, Daniela Rus, Alaa Maalouf

Computer Vision

MINERVA-Cultural: A Benchmark for Cultural and Multilingual Long Video Reasoning

Darshan Singh, Arsha Nagrani, Kawshik Manikantan et al. (+6 more)

Recent advancements in video models have shown tremendous progress, particularly in long video understanding. However, current benchmarks predominantly feature western-centric data and English as the dominant language, introducing significant biases in evaluation. To address this, we introduce MINERVA-Cultural, a challenging benchmark for multicultural and multilingual video reasoning. MINERVA-Cultural comprises high-quality, entirely human-generated annotations from diverse, region-specific cultural videos across 18 global locales. Unlike prior work that relies on automatic translations, MINERVA-Cultural provides complex questions, answers, and multi-step reasoning steps, all crafted in native languages. Making progress on MINERVA-Cultural requires a deeply situated understanding of visual cultural context. Furthermore, we leverage MINERVA-Cultural's reasoning traces to construct evidence-based graphs and propose a novel iterative strategy using these graphs to identify fine-grained errors in reasoning. Our evaluations reveal that SoTA Video-LLMs struggle significantly, performing substantially below human-level accuracy, with errors primarily stemming from the visual perception of cultural elements. MINERVA-Cultural will be publicly available under https://github.com/google-deepmind/neptune?tab=readme-ov-file\#minerva-cultural
Read the full abstract

All authors Darshan Singh, Arsha Nagrani, Kawshik Manikantan, Harman Singh, Dinesh Tewari, Tobias Weyand, Cordelia Schmid, Anelia Angelova, Shachi Dave

Computer Vision

CoMoVi: Co-Generation of 3D Human Motions and Realistic Videos

Chengfeng Zhao, Jiazhi Shu, Yubo Zhao et al. (+7 more)

In this paper, we find that the generation of 3D human motions and 2D human videos is intrinsically coupled. 3D motions provide the structural prior for plausibility and consistency in videos, while pre-trained video models offer strong generalization capabilities for motions. Based on this, we present CoMoVi, a co-generative framework that generates 3D human motions and videos synchronously within a single diffusion denoising loop. However, since the 3D human motions and the 2D human-centric videos have a modality gap between each other, we propose to project the 3D human motion into an effective 2D human motion representation that effectively aligns with the 2D videos. Then, we design a dual-branch diffusion model to couple human motion and the video generation process with mutual feature interaction and 3D-2D cross attentions. To train and evaluate our model, we curate CoMoVi-Dataset, a large-scale real-world human video dataset with text and motion annotations, covering diverse and challenging human motions. Extensive experiments demonstrate that our method generates high-quality 3D human motion with a better generalization ability and that our method can generate high-quality human-centric videos without external motion references.
Read the full abstract

All authors Chengfeng Zhao, Jiazhi Shu, Yubo Zhao, Tianyu Huang, Jiahao Lu, Zekai Gu, Chengwei Ren, Zhiyang Dou, Qing Shuai, Yuan Liu