Peer Revenue Impact factor: volatile

Papers

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

Computer Science Computer Vision

cs.CV Watch Computer Vision

Monday 27 July

Computer Vision

Rethinking Expert Training for Model Merging with Prompt Learning

Christos Georgakilas, Aniello Panariello, Samir El Karrat Moreno et al. (+3 more)

Model merging aims to combine multiple domain-specialized experts trained from a shared foundation model into a single multi-task model. Existing approaches largely focus on improving the merging procedure itself and typically assume experts obtained through full-parameter fine-tuning. In this work, we revisit expert training for model merging. We first show that prompt-based adaptation provides a strong baseline: independently learned prompts can be exploited across tasks while keeping the backbone fixed, avoiding the interference introduced by weight merging. Building on this observation, we introduce Dual-Tuned Experts (DTEs), a two-stage training strategy that first learns prompts and then fine-tunes the vision encoder. This reduces the magnitude of task-specific parameter updates and produces experts with higher merge compatibility. Experiments across multiple CLIP architectures, full fine-tuning, and LoRA experts show that DTEs consistently improve merged performance of standard merging approaches and remain effective even when combining heterogeneous sets of experts.
Read the full abstract

All authors Christos Georgakilas, Aniello Panariello, Samir El Karrat Moreno, Simone Calderara, Dimosthenis Karatzas, Joost van de Weijer

Computer Vision

ESRVS: Extreme Semi-Supervised Retinal Vessel Segmentation with a Single Annotated Image

Mingzhi Xu and Yizhe Zhang

Learning from minimal human supervision is a long-standing goal in medical image analysis, where dense expert annotations are costly. We study retinal vessel segmentation in an extreme semi-supervised setting with one annotated image and a pool of unlabeled images. We propose ESRVS, which selects a representative reference image for manual annotation and transfers vessel cues using target-domain-adapted DINOv3 features. ESRVS constructs a multi granular vessel prototype, combines prototype-similarity maps with a physics-inspired prior to generate initial pseudo-labels, and refines the transferred supervision through weighted pseudo-label training and adversarial refinement. Across eight public datasets, ESRVS achieves the best Dice and clDice on six datasets, and the best HD95 on all eight datasets among the compared semi-supervised methods, although those methods use 10 to 20% labeled data. With Mask2Former, ESRVS retains on average 93.7% of fully supervised Dice and 95.1% of fully supervised clDice. These results demonstrate the potential of foundation-model label propagation for highly label-efficient retinal vessel segmentation. Code is available at https://github.com/IAANNH/ESRVS.
Read the full abstract

Computer Vision

RP-OPSD: Resolution-Privileged On-Policy Self-Distillation for Multimodal Large Language Models

Qihui Zhu, Yuchen Wang, Zijian Wen et al. (+7 more)

On-Policy Self-Distillation (OPSD) uses privileged information available only to the teacher to provide dense token-level supervision on trajectories generated by the student. However, existing methods often rely on verified solution traces, explanations generated by external models, or manually localized visual evidence, which limits their scalable application to multimodal large language models. To address this issue, we exploit the information gap between high- and low-resolution views of the same image and propose RP-OPSD (Resolution-Privileged On-Policy Self-Distillation for Multimodal Large Language Models). During training, the student policy generates on-policy trajectories from images at one-quarter of the original resolution, while the teacher policy provides supervision using the original-resolution images. By minimizing the divergence between their output distributions along the student trajectories, the student learns the predictive behavior of the teacher under high-resolution inputs, thereby strengthening its low-resolution capability and transferring the learned improvement to original-resolution inference. RP-OPSD requires neither additional human annotations nor external models to generate solution traces but only image–question pairs. Experiments on Qwen3.5-9B show that RP-OPSD achieves a 5.45% relative improvement in average performance at the original resolution and a $1.78\times$ training speedup over OPSD. These results demonstrate that resolution differences can serve as a simple and scalable source of privileged information, providing an effective and efficient approach to on-policy self-distillation for multimodal large language models.
Read the full abstract

All authors Qihui Zhu, Yuchen Wang, Zijian Wen, Tao Zhang, Mengjie Zhang, Yang Liu, Shuangwu Chen, Siying Wu, Jian Yang, Xiaofeng Jiang

Computer Vision

Bigger or Cheaper? Scale and Quantization Effects on Uncertainty Signals in Vision-Language Models Under Image Degradation

M M Asif Ferdous

Vision-language models (VLMs) deployed on consumer hardware must decide when to answer and when to defer, and that decision depends on having a confidence signal that tracks correctness. A practitioner with a fixed memory budget faces a choice between a small model at full precision, the same small model quantized, and a larger model quantized into the same footprint – three configurations that push the confidence signal in opposing directions. We measure, on identical inputs, how model scale and 4-bit quantization affect two confidence signals in the Qwen2-VL family: the confidence a model states in natural language, and its own mean token probability over the answer it generates. Across 5,700 predictions spanning six realistic photographic degradations at three severities, we find that scale sharply improves the model's internal uncertainty signal (mean error-detection AUROC 0.80 to 0.98 from 2B to 7B) while its verbalized confidence stays weak and often at chance (mean 0.61 to 0.69): the gap between what the model knows and what it says widens rather than closes with size. We find that 4-bit quantization is nearly free for accuracy (-1.6 points) but expensive for the confidence signal (internal AUROC 0.95 to 0.80, and the verbalized-confidence parse rate collapses from 99% to 64%). For a fixed memory budget the recommendation is therefore to prefer a larger quantized model over a smaller full-precision one: 7B-4bit gives both the best accuracy and the best uncertainty signal (internal AUROC 0.98) of the three configurations that fit. We frame the results as selective-prediction operating points so they translate directly into a deployment recommendation, and we argue that error-detection AUROC, not calibration error, is the metric that exposes the difference between the two signals.
Read the full abstract

Computer Vision

MSVS-VAE: Multi-Scale Anchored VecSet for High-Fidelity 3D Reconstruction

Dehao Hao, Kaiyi Zhang, Tanghui Jia et al. (+10 more)

High-fidelity 3D generative modeling increasingly relies on the latent diffusion paradigm, where the reconstruction quality of the underlying 3D VAE becomes a primary bottleneck. Existing approaches largely follow two paradigms: sparse voxel-based representations achieve strong reconstruction quality but incur significant memory and computational overhead, while set-based representations are compact and continuous yet typically lag in fidelity due to latent sparsity and excessive global smoothness. We propose MSVS-VAE, a hierarchical set-based VAE that closes this fidelity gap without sacrificing compactness. Our key idea is to progressively densify anchored VecSet latents via hierarchical point-shuffle upsampling, increasing spatial capacity for fine-grained geometry modeling. To efficiently decode from the densified hierarchy, we replace global cross-attention with AVS-Conv, a geometry-aware local aggregation operator operating within local neighborhoods rather than the exhaustive latent set. We further introduce multi-scale query decoding to fuse coarse-to-fine latent features, where coarse scales provide stable global context, and fine scales refine localized geometry, reducing artifacts from overly local receptive fields. Extensive experiments on Objaverse, ABO, and in-the-wild benchmarks demonstrate that MSVS-VAE consistently outperforms prior set-based and voxel-based VAEs, delivering approximately 10x faster decoding than prior set-based methods and approximately 10x higher compactness than voxel-based baselines.
Read the full abstract

All authors Dehao Hao, Kaiyi Zhang, Tanghui Jia, Xiangjun Gao, Dongyu Yan, Weikai Chen, Zeyu Hu, Lingting Zhu, Yingda Yin, Runze Zhang, Li Yuan, Xin Wang, Long Quan

Computer Vision

InterOCF: Spatio-Temporal 2D-3D Interaction for Camera-Only 4D Occupancy Forecasting

Qi Zhang, Xinquan Yu, Kaiyi Zhang et al. (+1 more)

Camera-only 4D occupancy forecasting enables autonomous vehicles to predict future 3D semantic scenes solely from historical multi-view images, which is critical for driving safety. Even though current methods have achieved good performance, the strong spatial-temporal modeling between the input multi-view frames is still underexplored, which limits the performance of those methods in future 4D forecasting. To address this gap, we introduce a novel framework, InterOCF, for 4D occupancy forecasting that jointly models temporal dynamics in both 3D voxel-based representations and multi-view segmentation sequences, while explicitly incorporating feature interaction between the 2D and 3D branches. Our framework incorporates three core components: 1) A 3D Spatio-Temporal (3DST) module that learns volumetric dynamics from historical voxel states to predict future voxel states; 2) A 2D Spatio-Temporal (2DST) module employing an auxiliary multi-view temporal segmentation forecasting task to enhance temporal semantic dynamics; 3) A Spatio-Temporal Interaction Modeling (STIM) module that enables feature interaction between 2D and 3D representations. Experiments on the nuScenes, Lyft-Level5, and nuScenes-Occupancy datasets show that InterOCF consistently outperforms existing baseline approaches.
Read the full abstract

All authors Qi Zhang, Xinquan Yu, Kaiyi Zhang, Hui Huang

Computer Vision

MAViE: A Multi-scale Adaptive Vision Encoder for Fine-grained Visual Perception and Efficient Multimodal Reasoning

Shaofei Lei

Vision-language models commonly project all tokens produced by a pretrained vision encoder into a large language model. However, final-layer features can discard text, local attributes, and spatial relationships, while high-resolution inputs substantially increase context length and inference latency. We introduce \method, a Multi-scale Adaptive Vision Encoder. \method uses position-dependent gates to fuse shallow, intermediate, and deep features from a vision Transformer, preserving global semantics while enhancing edges, text, and local structure. It then performs question-conditioned token routing according to question relevance, local information content, global semantics, and spatial coverage, with a token budget that adapts to image complexity. To mitigate compression loss, we further introduce full-to-compressed representation distillation and a spatial diversity regularizer. In an illustrative simulation under a unified 7B language-model framework, \method reduces the average number of SigLIP-SO400M visual tokens from 729 to 146 (approximately 80.0%) and improves the mean score on VQAv2, GQA, TextVQA, ScienceQA-IMG, and MMBench by 2.2 percentage points, while reducing single-image time to first token from 228\,ms to 129\,ms. We provide the full model design and evaluation protocol. All reported numbers currently serve only as placeholders for paper organization and experimental design; formal claims require real training runs, independent replications, and official benchmark evaluation.
Read the full abstract

Computer Vision

IJCB-AFMFR 2026: Competition on Adapting Foundation Models for Face Recognition Using Synthetic Training Data

Tahar Chettaoui, Guray Ozgur, Eduarda Caldeira et al. (+13 more)

This paper presents a summary of the Competition on Adapting Foundation Models for Face Recognition Using Synthetic Training Data (AFMFR), held at the 2026 International Joint Conference on Biometrics (IJCB 2026). The competition received a total of eight valid submissions from four distinct teams across two complementary tracks: a Full Data Track, in which participants adapt the CLIP ViT-L/14 foundation model using large-scale synthetic identity data, and a Limited Data Track, designed to reflect more resource-constrained adaptation regimes. All training data was generated exclusively using IDPERTURB. Submitted solutions are ranked based on verification and identification performance across a diverse suite of benchmarks, including LFW, CFP-FP, AgeDB-30, CALFW, CPLFW, IJB-B, IJB-C, and TinyFace, using the Borda count method. Fairness evaluation is additionally conducted on the RFW dataset across four demographic groups. The results demonstrate that adaptation of the CLIP foundation model with synthetic training data substantially improves over the off-the-shelf model and, in several cases, surpasses the baseline. Notably, full fine-tuning with Sub-Center ArcFace (DMSTI-Neurotechnology) leads the Full Data Track, while rank-stabilized LoRA adaptation (Idiap-BSP) proves most effective under limited-data conditions.
Read the full abstract

All authors Tahar Chettaoui, Guray Ozgur, Eduarda Caldeira, Arturas Nakvosas, Hatef Otroshi Shahreza, Sébastien Marcel, Rishabh Shukla, Aditya Takkar, Rushil Khullar, Lalak Yadav, Gourav Gupta, Anant Gupta, Shiqi Yu, Vitomir Struc, Naser Damer, Fadi Boutros

Computer Vision

Accuracy potential of visual localization exploiting high-end street-level imagery

Jonas Meyer, Stephan Nebiker, Pascal Theiler et al. (+1 more)

Accurate and reliable pose information with respect to a reference frame is increasingly demanded across applications such as autonomous navigation, surveying, robotics, and augmented and mixed reality. Visual localization can serve as a complementary positioning modality to GNSS, whose applicability and accuracy are often limited. Yet, the accuracy potential of visual localization has not been systematically investigated against survey-grade demands. This is mainly due to the lack of publicly available, large-scale outdoor datasets with ground-truth poses in the sub-centimeter range. In this work, we address both gaps. We introduce a scalable visual localization pipeline that employs precisely georeferenced, high-resolution street-level imagery directly as the scene representation. It combines prior-guided reference candidate selection with on-the-fly local Structure-from-Motion reconstruction and PnP-based pose estimation. We further present the FHNW Muttenz dataset, a real-world dataset covering a contiguous 10 km street network mapped in two mobile mapping campaigns approximately 1.5 years apart. It consists of high-resolution reference imagery and query sequences acquired by four different cameras across five representative scenes. All images are precisely co-registered, yielding 6-DoF ground-truth poses in the sub-centimeter range. Using this dataset, we evaluate the accuracy potential of visual localization. Our experiments demonstrate median pose accuracies in the range of 1-5 cm for translation and 0.05-0.1{\deg} for rotation, reaching as low as 1 cm and 0.03{\deg} under favorable conditions. These results show that visual localization can complement survey-grade GNSS positioning, paving the way for 3D geospatial data acquisition using consumer devices and fully automated georeferencing approaches. The dataset is publicly available at: https://fhnw-muttenz-vl-dataset.github.io/.
Read the full abstract

All authors Jonas Meyer, Stephan Nebiker, Pascal Theiler, Norbert Haala

Computer Vision

Mixture-of-Thought-Tokens: Unifying Perception and Reasoning for Free-form Multimodal Grounding

Tianyi Gao, Han Fang, Tianyi Ding et al. (+9 more)

Multimodal Large Language Models have made great progress in grounding tasks, yet existing methods still struggle to unify precise localization and complex reasoning. For one thing, text-based methods rely on coordinates or index prediction, severely limiting the perceptual capabilities of the model for dense visual objects. Meanwhile, latent token-based methods employ special tokens without inherent spatial references and use a decoding mechanism that lacks thinking steps, weakening high-level reasoning capabilities. Consequently, developing a unified framework that excels in both perception and reasoning remains challenging. To address this, we propose Mixture-of-Thought-Tokens (Motto), a new free-form multimodal grounding method that bridges the perception-reasoning gap, enabling MLLMs to empower diverse, arbitrary grounding queries. Specifically, we introduce Spatially-Grounded Thought Tokenization to explicitly align special tokens with spatial locations for clear spatial correspondence and visual interpretability. We further design a Context-Adaptive Chain-of-Tokens that dynamically switch grounding modes within an interleaved reasoning chain, achieving robust grounding across tasks of varying complexity. In addition, we construct PR-Bench, a new referring expression comprehension benchmark to evaluate the perception-reasoning gap. Extensive experiments demonstrate that Motto achieves state-of-the-art performance across diverse free-form grounding tasks.
Read the full abstract

All authors Tianyi Gao, Han Fang, Tianyi Ding, Hao Li, Xin Wei, Hongbo Sun, Xiaodong Dong, Ye Yuan, Jinglin Xu, Kongming Liang, Hao Sun, Jingmin Xin

Computer Vision

GenSplatCodec: Feed-Forward Gaussian Splatting Compression via One-Step Diffusion

Qiang Hu, Zhenlong Wu, Lei Huang et al. (+3 more)

Feed-forward 3D Gaussian Splatting (3DGS) enables scalable scene reconstruction without per-scene optimization, yet produces dense Gaussians that are costly to store and transmit. Existing feed-forward Gaussian compression methods formulate decoding as deterministic representation recovery, which becomes inadequate at low bitrates when high-frequency textures and view-dependent appearance are discarded. Although generative models offer a promising alternative, using them as standalone post-processing decouples generation from the transmitted scene structure, thereby compromising cross-view consistency. To address these limitations, we propose GenSplatCodec, a unified feed-forward Gaussian codec that reformulates low-bitrate Gaussian compression as geometry-guided generative decoding. We present a detail-aware feed-forward Gaussian coding scheme within a dual-stream formulation, where the resulting compact Gaussian structural stream is complemented by a lightweight reference appearance stream. We further introduce a geometry-guided one-step generative decoding approach that jointly exploits decoded structural and appearance cues through hierarchical geometry control to reconstruct high-fidelity and view-consistent novel views. Finally, we develop a three-stage optimization strategy that stabilizes the learning of the unified codec and adapts the generative decoder to codec-derived structural and appearance cues. Extensive experiments across multiple datasets demonstrate that GenSplatCodec consistently achieves superior rate-distortion (RD) performance over existing methods.
Read the full abstract

All authors Qiang Hu, Zhenlong Wu, Lei Huang, Zihan Zheng, Xiaoyun Zhang, Wenjun Zhang

Computer Vision

Ambient pressure compensation and robust position control of oil-filled electric joint systems for underwater manipulators

Hongrui Wu, Songhui Wang and Xin Wang

Electric joint systems are significant elements of an underwater manipulator for its actuation, drive, and control. Working in an underwater environment, the joints suffer huge ambient pressure. To withstand it, the pressure compensation method is usually deployed, whereas the pressurized oil introduces sealing problems as well as parametric uncertainties and unknown disturbances for the dynamic model of the joint. To tackle these issues, this study proposes a design framework for the underwater oil-filled electric joint. The dynamics of the pressure compensation module is analyzed and the structure of the joint is optimized to seal the internal hydraulic oil. An uncertainty dynamic model of the oil-filled joint is established and a robust position controller is designed based on the structured singular value synthesis (mu-synthesis). Experimental results validate the feasibility of the proposed methods.
Read the full abstract

Computer Vision · Cross-listed from cs.LG

MXAttention: Data-Free Optimal Scaling and Pre-Normalization Quantization for MXFP4 Attention

Jianlin Yu, Jing Lin, Linghui Kong et al. (+13 more)

The quadratic cost of attention is a major bottleneck in diffusion-based video generation models. MXFP4 attention provides a promising path toward efficient inference, but direct MXFP4 quantization often degrades generation quality due to two numerical issues: the clipping-underflow trade-off from power-of-two scaling and the row-wise normalization error introduced in the softmax loop. We propose MXAttention, a data-free post-training quantization framework for MXFP4 attention. MXAttention introduces two components: Universal Optimal Scaling (UOS), which exploits the periodic structure of power-of-two microscaling to derive a distribution-independent optimal scaling boundary Qmax=7.25 without calibration or search, and Pre-Normalization Quantization (PNQ), which quantizes unnormalized softmax exponentials before row-wise summation to preserve normalization by construction. Experiments on Wan2.2 and HunyuanVideo show that MXAttention closes at least 95% of the VBench Imaging Quality gap between OCP MXFP4 and FP16, substantially improves frame-level similarity, and preserves FP16-level generation quality with less than 0.01 absolute degradation on all reported VBench metrics. MXAttention also achieves performance competitive with strong NVFP4-based baselines with negligible overhead when fused into the attention pipeline. The implementation is publicly available in MindIE-SD.
Read the full abstract

All authors Jianlin Yu, Jing Lin, Linghui Kong, Aiyue Chen, Weiyi Sun, Chenyu Zeng, Wangli Lan, Jinxi Li, Zhuo Zheng, Ziyang Yue, Danning Ke, Fei Yi, Tianchi Hu, Yuan Ding, Yiwu Yao, Junsong Wang

Computer Vision

HistoGPA: A Context-Conditioned Gene-Prior Attention Framework for Histology-Based Spatial Gene Expression Prediction

Ziang Liu, Xinhai Chen, Yigui Feng et al. (+3 more)

Predicting spatial gene expression from routine hematoxylin and eosin (H&E) images provides a practical complement to experimental spatial transcriptomics. Existing approaches focus on local or multi-scale visual features and often treat pretrained gene representations as fixed priors, although the interpretation of local morphology and the relevance of gene priors depend on tissue context. We propose HistoGPA, a context-conditioned gene-prior attention framework that uses a shared slide-level representation in two parallel pathways: one modulates local morphological features, whereas the other conditions pretrained gene embeddings and retrieves gene-prior information through cross-attention. This design enables each spatial location to retrieve context-adapted gene-prior information using its local morphology, position, and slide context. Across ten cancer types in HEST-1k, HistoGPA achieves the highest macro-averaged gene-wise Pearson correlation coefficient among the compared methods under the same evaluation protocol for both the top-50 and top-1,500 highly variable gene sets. Additional analyses show that HistoGPA better recovers the spatial expression patterns of cancer-associated genes and yields greater agreement between clusters derived independently from predicted and ground-truth expression profiles. Together, these findings motivate a context-dependent view of histology-to-expression prediction, in which local morphological representations and gene priors are jointly adapted to the broader tissue context.
Read the full abstract

All authors Ziang Liu, Xinhai Chen, Yigui Feng, Shuai Li, Qingyang Zhang, Jie Liu

Computer Vision

TaoMate: Anchor-Guided Memory Bridging Evolving and Reference States for Real-Time Audio-Video Digital Human Generation

Qijun Gan, Chenwei Zhang, Meiguang Jin et al. (+2 more)

Real-time long-form digital-human generation relies on causal models to extend audio-visual content while preserving subject appearance and audio-video synchronization across successive segments. A bounded cache retains local motion and phonetic context but discards older evidence, whereas attending to the complete generated history is computationally expensive and can propagate accumulated errors. We present \method, an anchor-guided persistent-memory framework for few-step joint audio-video generation. The framework preserves an immutable visual anchor, compresses completed video and audio blocks into fixed-capacity dynamic states, and retrieves those states through modality-specific residual attention without extending the active cache. A reference-aware modulation method additionally conditions video features on dynamic and anchor appearance statistics. Anchor-preserving causal-context distillation varies rollout horizon, prefix provenance, and cache-history reliability while keeping the immutable visual anchor unperturbed. By separating persistent memory from stage-local denoising dependencies, \method further admits stage-parallel execution across blocks, accelerating autoregressive inference without pipeline-specific retraining. We evaluate long-form video continuations with appearance, temporal, synchronization, facial, and speech diagnostics. Results show that \method preserves stable appearance across prompt-conditioned segments and strong audio-visual synchronization under autoregressive generation. Our project page is https://taoliveaigc.github.io/TaoMate.
Read the full abstract

All authors Qijun Gan, Chenwei Zhang, Meiguang Jin, Junfeng Ma, Qiu Shen

Computer Vision

PRISM: Prompt Refinement via Image-grounded Self-rewarding Mechanism for Text-to-Image Generation

Guo Tang, HongJie Luo, Tianxu Wang et al. (+2 more)

Text-to-image generation models can synthesize high-quality images from natural language descriptions, but their performance remains highly sensitive to prompt formulation. Existing prompt optimization methods mainly rely on text-side rewriting, prompt expansion, or external reward signals, offering limited image-grounded diagnosis and weak support for learning reusable optimisation policies. In this paper, we propose PRISM, a Prompt Refinement framework via Image-grounded Self-rewarding Mechanism. PRISM closes the prompt-image-feedback loop by interpreting generated images with structured visual diagnosis and scoring them along semantic consistency, aesthetic quality, and human preference alignment. It first initializes a unified VLM through multi-task supervised fine-tuning, and then improves the prompt policy via self-rewarding optimization with a hybrid ideal-point and Chebyshev reward. Extensive experiments show that PRISM improves holistic image quality and fine-grained semantic alignment, while providing interpretable feedback for targeted prompt refinement. The code is available at https://anonymous.4open.science/r/PRISM-FF81.
Read the full abstract

All authors Guo Tang, HongJie Luo, Tianxu Wang, Ying Zhang, Hao Wang

Computer Vision · Cross-listed from cs.IR

Energy Constrained Hierarchical Underwater Monitoring via Local Multi-Agent RAG

Mohamed Amine Janati, Laurent Gautier and Stéphane Barbot

Marine life monitoring is limited by strict energy constraints, poor underwater connectivity, and the high cost of transmitting raw multimodal data from remote deployments. This paper proposes a low-consumption underwater monitoring architecture that combines always-on edge sensing with selective high-performance local reasoning. The system follows a hierarchical master–satellite design in which ultra-low-power MAX78000/MAX78002 microcontrollers continuously monitor visual and acoustic signals, while an NVIDIA Jetson Orin NX is activated only for scheduled processing, event-driven analysis, or researcher interaction. Once active, the Jetson executes a fully local multimodal pipeline for data ingestion, visual target extraction, embedding-based indexing, species identification, retrieval-augmented reasoning, and automated reporting. BioCLIP/OpenCLIP embeddings are used to organize mission data, marine taxonomic references, scientific documents, and operational metadata in local ChromaDB collections. A dedicated identification layer combines visual similarity search, centroid-based classification, and supervised classifiers to support adaptive species recognition. A LangChain-based multi-agent framework coordinates query routing, structured analysis, energy management, hardware reconfiguration, and report generation. The architecture is evaluated through visual and acoustic monitoring case studies. The proposed system bridges ultra-low-power continuous sensing with local multimodal intelligence, enabling underwater stations to produce structured, researcher-ready knowledge while compressing local data for flexible acoustic, optical, or satellite transmission, minimizing both energy use and communication overhead.
Read the full abstract

Computer Vision

Multiview Multi-Person Human Mesh Recovery Under Large Scenes with Occlusions

Qi Zhang, Tao Yu, Jiechao He et al. (+2 more)

Human mesh recovery (HMR) aims to recover 3D human meshes from images. Most existing HMR benchmarks and methods focus on either multi-person reconstruction from a single view or single-person reconstruction from multiple views, where the number of subjects and the scene scale are relatively limited. Such settings are insufficient for real-world applications with large scenes and severe inter-person occlusions. To address this limitation, we introduce a large-scale synthetic benchmark for multiview multi-person HMR, termed MVMP-HMR. The proposed dataset contains 15 complex scenes with up to 50 camera views and 30 interacting persons, featuring large spatial coverage and severe occlusions, which significantly increases the difficulty of human mesh recovery. Based on this benchmark, we further propose a multiview multi-person whole-body human mesh recovery model, referred to as MVMP-HMR model. The model first fuses multiview features into a scene-level 3D feature volume, and then leverages pelvis joints predicted by a 3D pose estimation network to extract person-specific queries from the 3D feature volume. These human queries are cross-attended with the 3D feature volume and integrated to decode each person's 3D mesh. Moreover, we introduce two novel losses–the orientation loss and the 3D joint density loss–to alleviate orientation and pose ambiguities under severe occlusions. Experiments demonstrate that existing state-of-the-art HMR methods struggle on the proposed MVMP-HMR benchmark, while our method consistently outperforms prior SOTAs in large-scale scenes with severe occlusions.
Read the full abstract

All authors Qi Zhang, Tao Yu, Jiechao He, Antoni B. Chan, Hui Huang

Computer Vision

UMI3D: Robust 3D Generation on Unconstrained Multi-Image Inputs via Simultaneous Focus Cross-Attention Routing

Zefan Qu, Zhenwei Wang, Gerhard Petrus Hancke et al. (+1 more)

Recent 3D foundation models can generate high-quality assets from a single image, but degrade markedly on unconstrained multi-image inputs, often producing distorted geometry, over-smoothed textures, and chaotic colors. We argue that this failure stems not from limited model capacity, but from a mismatch between single-image cross-attention and the multi-image setting: existing models lack a principled way to decide which image each 3D voxel should trust at each denoising step. Revisiting recent single-image 3D foundation models, we show that explicitly routing each voxel to its most informative image is sufficient to unlock strong performance on inconsistent multi-image inputs. Based on this observation, we propose UMI3D, a training-free and plug-and-play framework that restructures cross-attention for unconstrained multi-image 3D generation. Its core, Simultaneous Focus Cross-Attention (SFC-Attn), activates all conditioning images at each denoising step while allowing each voxel to focus on the single image that best explains it. To enable this routing, we derive the Voxel Reference Score (VRS), a model-intrinsic metric for voxel–image affinity that requires no external matching, segmentation, or correspondence models. Extensive experiments show that UMI3D unlocks the multi-image potential of single-image 3D generation frameworks across diverse tasks. Project Page: UMI3D-Project.github.io.
Read the full abstract

All authors Zefan Qu, Zhenwei Wang, Gerhard Petrus Hancke, Rynson W.H. Lau

Computer Vision

Superpixel-Based QUBO for Scalable Quantum-Enhanced Medical Image Segmentation

Mohammad Chalhoub, Mahdi Chehimi, Laia Domingo et al. (+3 more)

Quadratic unconstrained binary optimization (QUBO) has emerged as a powerful framework for medical computing problems. Binary decision variables naturally represent clinical choices, making QUBO formulations well-suited for quantum annealing hardware. However, a fundamental scalability challenge limits practical deployment: problem size grows rapidly with input dimensionality, creating computational bottlenecks that restrict applications to simplified scenarios. This paper addresses this challenge through hierarchical problem reduction, as demonstrated in medical image segmentation, where pixel-level QUBO formulations create over 65,000 variables for a 256x256 image, forcing existing approaches to downsample to 42x42 resolution and discard 97% of pixel information. A superpixel-based QUBO framework is proposed using simple linear iterative clustering (SLIC) to group pixels into perceptually meaningful regions, then formulate segmentation as QUBO over a region adjacency graph (RAG) combining min-cut and smoothness objectives. Validation on INbreast mammography breast cancer images demonstrates a 4.2% improvement in segmentation quality (mean IoU 0.76 vs 0.73) with 33 computational speedup (0.67s vs 21.97s) and a 97.3% reduction in problem size (1764 to 48 variables), all achieved while processing full-resolution images rather than downsampled versions. The reduced problem size also fits well within current quantum annealer connectivity limits, removing the embedding overhead that has historically blocked direct deployment of pixel-level QUBO segmentation on quantum hardware.
Read the full abstract

All authors Mohammad Chalhoub, Mahdi Chehimi, Laia Domingo, Omar Alhussein, Ahmed Farouk, Saif Al-Kuwari

Computer Vision

SILICA: Repurposing Diffusion Priors for Joint Glass Segmentation and Depth Estimation

Tarun R, Anuj Verma, Laksh Nanwani et al. (+2 more)

Standard depth sensors systematically fail on transparent surfaces, creating corrupted 3D maps and severe navigation hazards. While specialized hardware sensors can detect glass, they lack modularity and have extensive hardware dependencies. Consequently, learning-based monocular depth estimation has emerged as a compelling alternative. However, domain-specific glass-aware monocular depth estimators struggle with unfamiliar indoor layouts; restricted by the severe scarcity of real-world glass depth annotations, they fail to generalize zero-shot to new settings. This motivates us to explore whether the extensive priors of text-to-image diffusion models can enable generalizable perception of transparent surfaces. We introduce SILICA, a unified pipeline leveraging these priors to jointly predict glass segmentation and glass-aware depth. This mutual information exchange establishes a robust spatial hierarchy, entirely eliminating the need for paired real-world glass depth annotations. Subsequently, we use the predicted segmentation mask to explicitly filter incorrect glass depth points from standard sensors, recovering accurate metric glass depth for downstream 3D mapping and autonomous collision avoidance. Supported by our novel Mirage 18k dataset, extensive experiments demonstrate that SILICA achieves remarkable zero-shot transfer across diverse, unseen environments, outperforming state-of-the-art models by almost 20% and setting a new benchmark for transparent surface perception.
Read the full abstract

All authors Tarun R, Anuj Verma, Laksh Nanwani, Sourav Garg, K. Madhava Krishna

Computer Vision

FilmBench: A Film-Grade Benchmark for Cinematic Video Generation

Shengyi Wang, Niantong Li, Guangzheng Hu et al. (+27 more)

Progress in video generation keeps narrowing the visual gap between AI-generated and professionally produced footage, yet most benchmarks still draw prompts from web sources or LLM templates and score them with untrained, generic multimodal models. More fundamentally, their evaluation taxonomies remain rudimentary (overall visual quality, coarse text alignment and temporal smoothness) rather than the professional Cinematic Language criteria by which films are actually made and judged, so they assess basic video plausibility rather than film-grade craft. We introduce FilmBench, a text-to-video (T2V) and reference-to-video (R2V) benchmark grounded in the professional Cinematic Language of the film- academy tradition and co-developed with directors and faculty from the Beijing Film Academy and the Hujing Digital Media & Entertainment Group film studio. It rests on three choices. First, prompts are reverse-engineered from clips of award-winning films spanning 20 cinematic genres and chosen by professional directors, so every prompt is anchored to a verified live-action reference; the prompts follow real shot lists, and most script multiple shots (1,056 of the 1,169 prompts are multi-shot), unlike prior single-clip benchmarks. Second, evaluation follows a three-level Cinematic taxonomy of 3 axes, 12 components and 35 (T2V) +3 (R2V-only) sub-metrics. Third, we develop an in-house expert-grade automatic evaluation agent and open-source its core suite of Cinematic Language operators (FilmOps). Benchmarking leading video generation models (9 for T2V, 7 for R2V), the evaluator reproduces the human model ranking at model-level Spearman h̊o} = 0.95 (T2V) and 0.96 (R2V). Scores fall well below prior web-style benchmarks, with two consistent gaps in dynamic aesthetics and a marked single- to multi-shot performance drop that widens for weaker models.
Read the full abstract

All authors Shengyi Wang, Niantong Li, Guangzheng Hu, Hong Qi, Fei Ding, Weixu Qiao, Jinlin Wang, Xiaotong Lv, Peng Han, Zimeng Li, Fanshu Ding, Yushu Wang, Han Wu, Jingjing Chen, Chongxiao Wang, Yanhao Wu, Chenglong Huang, Xiaoqian Zhu, Jie Tian, Hua Li, Jingjing Fan, Mingshuang Tang, Zhong Li, Hengxia Qiang, Weibin Chen, Jinyang Zhen, Bing Zhao, Lin Qu, Jing Li, Hu Wei

Computer Vision

MATS: A novel multi-modality multi-task learning framework for 3D perception in autonomous driving

Junchen Huo, Wanming Hao, Song Wang et al. (+3 more)

Multi-modality data from different sensors provides rich complementary information for 3D perception, becoming an essential component in reliable autonomous driving systems. Current research typically designs intricate and complex fusion strategies to integrate information from multimodal data on a unified bird's-eye-view (BEV) feature map for the joint learning of multiple perception tasks. However, such a single feature map hardly carries sufficient information to simultaneously meet the requirements of various perception tasks, leading to a very limited perception performance. To mitigate this limitation, this paper proposes MATS, a novel multi-modality multi-task learning approach with modality-adaptive BEV fusion and task-specific Mixture-of-Experts (MoE) for 3D perception. Specifically, a simple modality-adaptive BEV fusion module is designed to adaptively recalibrate the BEV features by modeling the global cross-modality dependencies, generating diverse BEV feature maps for various perception tasks. For joint multi-task learning, this paper proposes a task-specific MoE module to decouple the tasks and enable the network to automatically choose the appropriate BEV feature candidates for each specific task. To validate the effectiveness of the proposed approach, we conduct extensive experiments on the large-scale benchmark nuScenes. With the camera- and LiDAR-modality input data, the proposed approach outperforms the state-of-the-art (SOTA) by a significant margin. Furthermore, the experimental results on the single tasks show that the proposed approach significantly outperforms the baselines. The code and trained models will be available upon publication.
Read the full abstract

All authors Junchen Huo, Wanming Hao, Song Wang, Enqing Chen, Shouyi Yang, Guanghui Wang

Computer Vision

TreeAdapter: Hierarchical Taxonomy-Guided Adapter Composition for Fine-Grained Species Image Generation

Yuze Sun, Zhongjie Duan and Yingda Chen

Although general text-to-image models excel in open-domain generation, their performance degrades significantly in specialized downstream domains, particularly when generating images of rare biological species. Hindered by long-tailed distributions, general models struggle to capture subtle fine-grained details, while per-species fine-tuning methods over-isolate individual species and consequently ignore the shared visual features among closely related taxa. To address this, we propose TreeAdapter, a novel framework that explicitly leverages hierarchical taxonomic data. Rather than using a monolithic model or independent per-species modules, TreeAdapter attaches lightweight adapters to every node of the taxonomic tree. Specifically, leaf-node adapters capture species-specific visual traits, while internal-node adapters encapsulate shared semantics among descendant taxa. We introduce a two-stage training paradigm where ancestor adapters are optimized to model only the residual visual features unexplained by their descendants. This model architecture and training paradigm enable the model to fully leverage hierarchical information, ensuring the accurate generation of visual features for each species. Extensive experiments across three large-scale biodiversity benchmarks demonstrate that TreeAdapter achieves state-of-the-art fine-grained generation quality, outperforming both general-purpose and domain-specific baselines.
Read the full abstract

Computer Vision

Effect of User-Prompted Priors on Semi-Automated Cancer Lesion Segmentation in Whole-Body Computed Tomography

Isac Stark, Johan Öfverstedt, Elin Lundström et al. (+3 more)

In clinical oncology studies, metastatic cancer is commonly evaluated using "Response Evaluation Criteria in Solid Tumors" (RECIST), in which the diameter of up to five lesions is measured and followed over the course of treatment. However, RECIST shows limited correlation with overall survival. Total tumour volume (TTV) is a stronger predictor but typically relies on manual ground-truth segmentation of all lesions, which is time-consuming and requires expert domain knowledge. Semi-automated approaches leveraging user-prompted priors, such as bounding boxes and single-slice contours, as inputs to automated segmentation methods can facilitate the generation of ground-truth segmentations. This work investigates the impact of different user-prompted priors on semi-automated cancer lesion segmentation performance in whole-body computed tomography. Across 3-fold cross-validation and external testing, more complex spatial priors consistently improved performance, with contour priors from three orthogonal planes (axial, coronal and sagittal) achieving the best results. On the external test (n=3865 lesions), this approach achieved a mean Dice score of 0.882, compared to a mean Dice score of 0.671 for the baseline model with no spatial prior. These findings suggest that the use of multi-plane orthogonal user-prompted priors can improve semi-automated tumour lesion segmentation and support efficient generation of high-quality volumetric ground-truth data.
Read the full abstract

All authors Isac Stark, Johan Öfverstedt, Elin Lundström, Simon Ekström, Håkan Ahlström, Joel Kullberg

Computer Vision · Cross-listed from cs.RO

Surgical Re-enactment for Operating Room Workflow Datasets

Jana Nina Friedrich, Andrea Karin Maria Ross, Angelo Henriques et al. (+3 more)

The introduction of new technologies, such as surgical robots, is driving the vision of a connected, smart operating room (OR). However, realizing this vision requires a deep understanding of surgical workflows, which relies on realistic datasets capturing the actions of all OR personnel from both full room and surgical field perspectives. Acquiring such data in real ORs is prohibitively challenging due to factors such as ethics committee approvals, limited space for camera installation, and sterility regulations preventing the use of tracking markers. We present a step-by-step methodology for re-enacting complete surgical procedures in a reconstructed OR. This approach enables the creation of repeatable and annotatable workflow datasets for training activity recognition models, generating scene graphs, and formalizing surgical process models. Developed for robot-assisted ophthalmic surgery, our methodology combines expert consultation, structured workflow formalization, OR reconstruction, role-based training, real OR observation, and iterative recording with post-take debriefing. We provide concrete recommendations to allow other research groups to seamlessly adopt this methodology for their own surgical domains.
Read the full abstract

All authors Jana Nina Friedrich, Andrea Karin Maria Ross, Angelo Henriques, Mario Peter Martin Weisser, Ling Zhang, Mohammad Ali Nasseri

Computer Vision

Reasoning to Regulate: Chain-of-Thought for Traffic Rule Understanding

Yueru Luo, Xu Yan, Changqing Zhou et al. (+5 more)

Understanding and complying with traffic regulations is a safety-critical requirement for autonomous driving, yet remains challenging due to the diversity and context dependence of traffic signage. Importantly, regulation understanding is not a simple recognition task, but a reasoning problem: whether a rule applies depends on interpreting the sign in relation to the spatial layout of lanes and scene context. To support such reasoning, MapDR provide fine-grained annotations that link each traffic sign's regulatory rules to the specific lanes they govern. Existing methods, however, largely treat this as direct sequence prediction, ignoring the underlying reasoning that connects sign semantics and map structure. To address this limitation, we explicitly incorporate reasoning into this task and propose a framework that equips vision-language models (VLMs) with chain-of-thought (CoT) capabilities. We first design a scalable CoT curation pipeline that bootstraps rationales from a strong LLM through a two-round strategy and employs a VLM-based verifier to filter out incorrect cases, yielding a high-quality set of (CoT, answer) pairs. Building on this foundation, we adopt a two-stage training scheme: supervised fine-tuning (SFT) to teach rationale-to-answer generation, followed by GRPO reinforcement learning with answer-grounded, fine-grained rewards to further improve final answer accuracy. Extensive experiments on MapDR show that our approach significantly improves both interpretability and accuracy, establishing the first reasoning-based framework for regulation-aware autonomous driving.
Read the full abstract

All authors Yueru Luo, Xu Yan, Changqing Zhou, Yiming Yang, Chao Zhan, Shuqi Mei, Chao Zheng, Zhen Li

Computer Vision

Face Age Verification Vulnerabilities Under Simple Appearance Manipulations

Ioannis Sarridis, Ioannis Kompatsiaris and Symeon Papadopoulos

Online platforms increasingly rely on automated age estimation systems to enforce minimum-age policies. Focusing on vision-based models designed for this task, concerns arise regarding their robustness to simple appearance changes that underage individuals may use to bypass such systems, such as drawing a mustache or applying lipstick. In this work, we present a systematic study of age verification robustness by simulating visual alterations that can be easily achieved by underage individuals. We evaluate seven models, including vision, vision-language, and multimodal large language models, across three datasets and four manipulation types. Interestingly, under drawn beard stubble, up to 61% of True Negatives are flipped into False Positives. Furthermore, we investigate how different demographics are affected by such manipulations, finding that Indians are more affected by beard stubble manipulations, while females are more affected than males across all manipulations. Finally, we explore how these biases can be mitigated using bias mitigation methodologies in lightweight linear probe settings.
Read the full abstract

Computer Vision

LCMamNet: A Lightweight Cross-scale Mamba Network for Infrared Small Target Detection

Yuhao Fan, Le Hui and Yuchao Dai

Infrared small target detection (IRSTD) is important for low-altitude perception, unmanned-system warning, and security monitoring. However, weak targets in infrared imagery usually occupy only a few pixels and are easily submerged by cloud clutter, ground edges, and bright noise, making it difficult for lightweight segmentation-based methods to preserve local target structures while suppressing background interference. To address these challenges, we propose LCMamNet, a lightweight cross-scale Mamba network that progressively enhances local target structures, interacts cross-scale context in a latent space, and restores spatial details with background suppression. Specifically, a compact hierarchical encoder with cross-shaped directional bottleneck residual (CDBR) blocks strengthens direction-sensitive target structures under a small computation budget. A latent dense cross-scale fusion (LDCF) module then performs dense all-level interaction through bidirectional Mamba modeling and reorganizes the interacted features into stable hierarchical semantics. Finally, a progressive decoder selectively recovers shallow spatial details while suppressing irrelevant background textures. Extensive experiments on IRSTD-1k, NUAA-SIRST, and NUDT-SIRST show that the proposed network achieves mIoU scores of 71.25%, 79.60%, and 95.58%, respectively, with only 1.175M parameters and 6.91 GFLOPs. It also runs with a mean inference latency of 6.62 ms, and deployment results on an NVIDIA Jetson Orin NX 16G SUPER further demonstrate its practical potential for real-time edge inference. The code and checkpoints are publicly available at https://github.com/Haoyu096/LCMamNet.
Read the full abstract

Computer Vision

Do Current Retrievers Cover All the Evidence? A Controlled Study of Conjunctive Cross-Page Retrieval

Sungguk Cha, DongWook Kim, Mintae Kim et al. (+3 more)

Finding a long document relevant to a multi-part request is not the same as establishing that it contains every requested piece of evidence. We study this gap for conjunctive document retrieval, where two or three explicit conditions must be supported on different pages of one document. We use n-Clue as a controlled measurement instrument: 1{,}000 queries over 2{,}021 documents pair all-condition golds with naturally occurring documents that satisfy only a subset, and a complete-first success requires a top-10 gold to precede every released subset qrel. Across 70 configurations, condition-wise decomposition improves two dense backbones by 6.8–7.3 points and lexical–visual fusion adds 8.7, while four generic rerankers all reduce Gold-NDCG; these directions replicate on a four-source stress set. Scaling one dense family from 0.6B to 8B changes complete-first success by 0.0 points. The strongest displayed hybrid illustrates the resulting gap: it finds a gold for 81.1% of queries but succeeds complete-first on only 35.8%, and the gap persists across condition count, target length, candidate density, query rendering, and the four-source stress set. Finally, page-aware visual systems surface stored support for every condition on only 5.1–5.3% of queries. These results identify condition coverage, rather than gold discovery alone, as the central bottleneck.
Read the full abstract

All authors Sungguk Cha, DongWook Kim, Mintae Kim, Youngsub Han, Byoung-Ki Jeon, Sangyeob Lee