cs.CV · 2026-06-02 · No. 13

Computer Vision and Pattern Recognition, 2026-06-02.

29 new papers in cs.CV. Titles, authors, abstracts. Links to arXiv. Want this in your inbox every morning? Subscribe →

01 — The papers

29 entries
  1. 01

    Mitigating Perceptual Judgment Bias in Multimodal LLM-as-a-Judge via Perceptual Perturbation and Reward Modeling

    Seojeong Park, Jiho Choi, Junyong Kang, Seonho Lee, Jaeyo Shin, Hyunjung Shim

    cs.CV · cs.AI

    Recent multimodal large language models have demonstrated strong reasoning ability, yet their reliability as automated evaluators remains limited by a critical weakness: when visual evidence conflicts with textual cues, MLLM judges tend to reward plausible narratives over perceptually correct answers. We identify and systematically analyze this phenomenon, which we term Perceptual Judgment Bias. Through controlled visual perturbations,...

    arxiv.org/abs/2606.02578 · PDF

  2. 02

    ProtoAda: Prototype-Guided Adaptive Adapter Expansion and Geometric Consolidation for Multimodal Continual Instruction Tuning

    Yu-Cheng Shi, Zhen-Hao Xie, Jun-Tao Tang, Da-Wei Zhou

    cs.CV · cs.LG

    Multimodal Large Language Models (MLLMs) achieve strong performance through instruction tuning, but real-world deployment requires them to continually acquire new vision-language capabilities, making Multimodal Continual Instruction Tuning (MCIT) essential. To reduce inter-task interference and promote collaboration, recent methods often employ sparse architectures like Mixture of LoRA Experts with image-text similarity routing. However,...

    arxiv.org/abs/2606.02576 · PDF

  3. 03

    AdaCodec: A Predictive Visual Code for Video MLLMs

    Haowen Hou, Zhen Huang, Zheming Liang, Qingyi Si, Chenglin Li, Shuai Dong, Kele Shao, Ruilin Li, Dianyi Wang, Nan...

    cs.CV · cs.AI · cs.CL

    Video is temporally redundant: adjacent frames usually share most objects, background, and layout. Yet existing video multimodal large language models (video MLLMs) usually encode each sampled frame as an independent RGB image, causing visual tokens to repeat content already present in earlier frames. This suggests a more direct video interface: send a full reference frame only when the scene cannot be predicted well from prior context, and...

    arxiv.org/abs/2606.02569 · PDF

  4. 04

    Modeling Depth Ambiguity: A Mixture-Density Representation for Flying-Point-Free Depth Estimation

    Siyuan Bian, Congrong Xu, Jun Gao

    cs.CV · cs.AI

    Despite advances in depth estimation, flying points remain a persistent failure mode: near object boundaries, depth estimators often predict spurious 3D points in the empty space between foreground and background surfaces. We trace this artifact to a standard modeling choice: assigning each pixel a single depth hypothesis. At boundaries, a pixel can straddle a foreground and a background surface, so its true depth is ambiguous between the...

    arxiv.org/abs/2606.02552 · PDF

  5. 05

    Why Not Hyperparameter-Friendly Optimisation? A Monotonic Adaptive Norm Rescaling Approach For Long-Tailed Recognition

    Shuo Zhang, Chenqi Li, Tingting Zhu

    cs.CV · cs.AI

    Long-tailed recognition poses a significant challenge for deep learning. The two-stage decoupling paradigm, which separates representation learning from classifier retraining, offers a promising solution. During the classifier retraining stage, adaptive norm rescaling is a popular technique. It adjusts the per-class weight norms via parameter regularization, which inevitably introduces hyperparameters. However, many studies report that...

    arxiv.org/abs/2606.02526 · PDF

  6. 06

    Moment-Video: Diagnosing Temporal Fidelity of Video MLLMs on Momentary Visual Events

    Xiaolin Liu, Yilun Zhu, Xiangyu Zhao, Xuehui Wang, Yan Li, Xin Li, Haoyu Cao, Xing Sun, Shaofeng Zhang, Xu Yang,...

    cs.CV · cs.AI

    Video multimodal large language models (MLLMs) have made rapid progress on general and long-form video understanding, yet their ability to preserve brief answer-critical visual evidence remains underexplored. Many practical questions are determined by momentary visual events: localized actions or state transitions that may last only a few frames. Such evidence can be skipped by sparse frame sampling, suppressed by visual-token compression, or...

    arxiv.org/abs/2606.02522 · PDF

  7. 07

    MASER: Modality-Adaptive Specialist Routing for Embodied 3D Spatial Intelligence

    Hilton Raj, Vishnuram AV

    cs.CV · cs.AI

    In 3D environments, Embodied Agents answer spatially relevant questions through reasoning from a mixture of modalities including natural language, RGB images, point clouds, depth maps and camera poses. Existing Vision-Language models (VLMs) are fine-tuned over a single modality. This completely ignores the question semantics which may favor a different modality than the finetuned modality. To address this, we propose MASER (Modality-Adaptive...

    arxiv.org/abs/2606.02463 · PDF

  8. 08

    Initialization is Half the Battle: Generating Diverse Images from a Guidance Potential Posterior

    Xiang Li, Dianbo Liu, Kenji Kawaguchi

    cs.CV · cs.AI

    Despite the remarkable fidelity of generative models, they frequently suffer from mode collapse. Existing strategies for enhancing diversity predominantly focus on intervening during the generation trajectory. We identify a critical oversight that the standard Gaussian initialization often causes trajectories to collapse into dominant modes because it is agnostic to the guidance potential landscape. In this work, we formulate selecting the...

    arxiv.org/abs/2606.02453 · PDF

  9. 09

    GC-MoE: Genomics-Guided Cell-Type-Specific Mixture of Experts for Histology-Based Single-Cell Spatial Transcriptomics

    Kaito Shiku, Ahtisham Fazeel Abbasi, Ryoma Bise, Yuichiro Iwashita, Kazuya Nishimura, Andreas Dengel, Muhammad Nabeel Asim

    cs.CV · cs.AI · cs.LG

    Histology-based single-cell spatial transcriptomics (ST) estimation aims to predict gene expression for individual cells from histopathological images and cell locations, reducing the need for costly single-cell ST measurements. Unlike existing histology-to-ST methods that mainly predict spot-level profiles for local regions containing multiple cells, this task requires modeling cell-to-cell expression variability, which is strongly...

    arxiv.org/abs/2606.02424 · PDF

  10. 10

    Do Multimodal Agents Really Benefit from Tool Use? A Systematic Study of Capability Gains

    Garvin Guo, Donglei Yu, Yu Chen, Xiang Wang, Shuai Li, Xinpei Zhao, Huaxing Liu, Qinghao Wang, Minpeng Liao

    cs.CV · cs.AI

    Tool-augmented multimodal agents show strong benchmark gains, often taken as evidence that agents have learned to use tools. We argue that this interpretation can be premature: a tool-call trace alone does not show whether the tool supplied answer-critical information. We study two representative ``thinking with images'' agents, Thyme and DeepEyesV2, across real-world understanding, OCR, chart understanding, and mathematical reasoning. Each...

    arxiv.org/abs/2606.02357 · PDF

  11. 11

    Hallucination-Aware Diffusion Sampling for Inverse Problems via Robust Prior Updates

    Pengfei Jin, Yiqi Tian, Kailong Fan, Bingjie Qi, Quanzheng Li

    cs.CV · cs.LG

    Diffusion-based inverse problem solvers can produce realistic reconstructions, but realism alone does not ensure that the recovered details are supported by the measurement. We study this failure as measurement-conditioned hallucination: visually meaningful content that is either implausible or inconsistent with the measured instance. Our analysis separates Bayes-rule-based diffusion inverse solvers into a prior update and a...

    arxiv.org/abs/2606.02331 · PDF

  12. 12

    Deep Learning for Remote Sensing to Improve Flood Inundation Mapping

    Yogesh Bhattarai, Vijay Chaudhary, Wai Lim Kim, Sanjib Sharma

    cs.CV · cs.LG

    Flooding is the most pervasive natural disaster worldwide. Timely and accurate flood inundation mapping are essential for informing disaster risk management. Optical satellite missions provide high-resolution, multispectral observations critical for flood detection and inundation mapping. However, their operational utility is severely constrained by cloud cover during extreme precipitation events. Conventional cloud-removal techniques based...

    arxiv.org/abs/2606.02310 · PDF

  13. 13

    Cross-modal linkage risk in clinical vision-language models

    Soroosh Tayebi Arasteh, Mahshad Lotfinia, Sven Nebelung, Daniel Truhn

    cs.CV · cs.AI · cs.CL · cs.LG

    Vision-language models (VLMs) trained on paired chest radiographs and radiology reports learn a shared embedding space that can preserve instance-level image-report correspondence. This poses a privacy risk in settings where radiographs and reports are deliberately kept separate after acquisition, such as image-only data sharing or access-controlled reports, because a de-identified image may be re-linked to its original narrative report...

    arxiv.org/abs/2606.02276 · PDF

  14. 14

    Towards Resolving Optimization Conflicts Between Image- and Text-Based Person Re-Identification

    Karina Kvanchiani, Timur Mamedov

    cs.CV · cs.AI · cs.LG

    The joint optimization of image-based (I2I) and text-based (T2I) person re-identification (ReID) is hindered by modality discrepancies and conflicting training objectives, leading to suboptimal shared representations. While I2I ReID focuses on identity-level invariance across images of the same person, T2I ReID is driven by instance-specific textual descriptions tied to unique visual traits. This paper explores the fundamental difference...

    arxiv.org/abs/2606.02242 · PDF

  15. 15

    CORE-MTL: Rethinking Gradient Balancing via Causal Orthogonal Representations

    Chengfeng Wu, Tao Zou, Yanru Wu, Jingge Wang

    cs.CV · cs.LG

    Multi-task learning (MTL) aims to construct a joint model for multiple tasks by sharing a common representation across domains. To achieve this goal, existing optimization-centric methods either balance task gradients or modify the shared architecture. However, as these approaches remain agnostic to the content of the shared representation, they fail to disentangle task-relevant structure from spurious context, leading to negative transfer...

    arxiv.org/abs/2606.02221 · PDF

  16. 16

    Order within Chaos: Capturing Intrinsic Energy Anomalies for AI-Manipulated Image Forgery Localization

    Yiming Wang, Baiqi Wu, Qingming Li, Jiahao Chen, Tong Zhang, Shouling Ji

    cs.CV · cs.AI

    Recent advancements in generative AI have led to image editing models capable of producing realistic forgeries that evade traditional image forgery localization methods, as these approaches depend on physical noise absent in synthetic data. To address this challenge, we theoretically demonstrate that the diffusion process inherently suppresses local high-frequency variance, creating a statistical energy gap that is distinguishable from the...

    arxiv.org/abs/2606.02178 · PDF

  17. 17

    Disentanglement-Based Equivariant Learning for Compositional VQA

    Zhou Du, Zhaoquan Yuan, Xiao Wu, Changsheng Xu

    cs.CV · cs.LG

    Compositional visual question answering (VQA) represents a challenging yet fundamental task that requires models to comprehend novel combinations of previously learned concepts. The current methods often overlook the disentanglement of underlying concepts and are restricted in terms of their ability to effectively capture the compositional variation mechanism. Moreover, the state-of-the-art techniques depend on additional clues for training,...

    arxiv.org/abs/2606.02168 · PDF

  18. 18

    Multimodal Approaches for Visually-Rich Document Type Classification: A Comparative Analysis

    Catyana Heyne, Jürgen Frikel, Filippo Riccio

    cs.CV · cs.AI · cs.CL · cs.IR

    Document type classification in visually rich documents remains challenging, as relevant information is distributed across textual, visual, and layout modalities. To capture this complexity, current approaches rely on diverse multimodal modeling strategies, resulting in heterogeneous architectures that complicate systematic comparison. This variability is also reflected in existing comparative studies, which often rely on heterogeneous...

    arxiv.org/abs/2606.02162 · PDF

  19. 19

    Understanding-Enhanced Model Collaboration for Long-Tailed Egocentric Mistake Detection

    Boyu Han, Qianqian Xu, Shilong Bao, Zhiyong Yang, Ruochen Cui, Qingming Huang

    cs.CV · cs.AI · cs.LG

    In this report, we address the problem of determining whether a user performs an action incorrectly from egocentric video data. To this end, we propose an Understanding-Enhanced Model Collaboration Method (UE-MCM) that combines efficient coarse-grained video understanding with accurate fine-grained action reasoning. Specifically, UE-MCM contains a small model branch and a large model branch. The large model branch focuses on whether the...

    arxiv.org/abs/2606.02120 · PDF

  20. 20

    Jailbreaking Multimodal Large Language Models using Multi-Clip Video

    Choongwon Kang, Seungjong Sun, Hyunmin Jun, Jang Hyun Kim

    cs.CV · cs.AI · cs.CL

    As multimodal large language models (MLLMs) have advanced to process video inputs, concerns have emerged about their potential for malicious misuse. Prior jailbreak studies have shown that safety alignment in MLLMs can be bypassed through visual inputs, yet it remains unclear which properties of video inputs induce this vulnerability. To address this gap, we introduce Multi-Clip Video (MCV) SafetyBench, a dataset of 2,920 videos designed to...

    arxiv.org/abs/2606.02111 · PDF

  21. 21

    Fast and Lightweight Novel View Synthesis with Differentiable Multiplane Image

    Kaidi Zhang, Guanxu Zhu

    cs.CV · cs.AI

    Recently, novel view synthesis has witnessed remarkable progress, with mainstream methods such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) delivering impressive results. However, these approaches often struggle to balance rendering speed and model size, and their optimization-based training can be highly time-consuming. Furthermore, they typically rely on dense observations, often failing to produce satisfactory results...

    arxiv.org/abs/2606.02068 · PDF

  22. 22

    Attention mechanisms and transfer learning for robust peach leaf damage classification under domain shift

    Adrián Cánovas-Rodriguez, Miguel A. González-Illán, Maria Fernanda García-Cruz, Pedro Nortes Tortosa, José Salvador...

    cs.CV · cs.AI

    Artificial intelligence provides a practical framework for crop damage assessment from imagery data, supporting early decision-making in agricultural management. In peach orchards, climate change increases abiotic stress and biotic pressures, including pests and diseases, which often produce visually similar foliar symptoms. This overlap makes manual diagnosis difficult, especially across multiple fields with varying environmental conditions,...

    arxiv.org/abs/2606.02045 · PDF

  23. 23

    Ranking vs. Assignment: The Metric Mismatch in Multi-View Object Association

    Matvei Shelukhan, Timur Mamedov, Aleksandr Chukhrov, Karina Kvanchiani

    cs.CV · cs.AI · cs.LG

    Multi-view object association is an important computer vision problem that underlies many multi-camera perception tasks. While this task is naturally formulated as a constrained one-to-one matching problem, recent works heavily rely on pairwise ranking metrics like AP and FPR-95 for model evaluation. We highlight a fundamental mismatch between these metrics and the actual assignment objective. Theoretically, we show that AP and FPR-95 can be...

    arxiv.org/abs/2606.02022 · PDF

  24. 24

    Towards 3D-Aware Video Diffusion Models: Render-Free Human Motion Control with Mesh Tokenization

    Jingyun Liang, Min Wei, Shikai Li, Yizeng Han, Hangjie Yuan, Lei Sun, Weihua Chen, Fan Wang

    cs.CV · cs.AI · eess.IV

    Diffusion models have shown remarkable success in video generation. However, whether such models are truly aware of the 3D structure underlying visual observations, rather than simply reproducing plausible 2D projections, remains an open question. In this work, we investigate this question through human motion control, a task that requires precise modelling of 3D human geometry, motion, camera viewpoint, and scene context. Unlike prior...

    arxiv.org/abs/2606.02000 · PDF

  25. 25

    A Structured Benchmark for Text-Guided Anomaly Detection: When Language Stops Conditioning the Decision

    Stefano Samele, Eugenio Lomurno, Teodora Jovanovic, Sanjay Shivakumar Manohar, Alberto Crivellaro, Matteo Matteucci

    cs.CV · cs.AI · cs.LG

    Industrial anomaly detection has historically been a unimodal task. Recent multimodal vision-language models have produced systems that admit textual input alongside the image and are presented as enabling text-guided zero- and few-shot inspection. Yet these methods are evaluated with protocols inherited from unimodal benchmarks that hold the textual condition constant and therefore cannot measure whether language conditions the decision;...

    arxiv.org/abs/2606.01992 · PDF

  26. 26

    Parameter-Efficient Fine-Tuning of Large Pretrained Models for Instance Segmentation Tasks

    Nermeen Abou Baker, David Rohrschneider, Uwe Handmann

    cs.CV · cs.AI

    Research and applications in artificial intelligence have recently shifted with the rise of large pretrained models, which deliver state-of-the-art results across numerous tasks. However, the substantial increase in parameters introduces a need for parameter-efficient training strategies. Despite significant advancements, limited research has explored parameter-efficient fine-tuning (PEFT) methods in the context of transformer-based models...

    arxiv.org/abs/2606.01947 · PDF

  27. 27

    The Image Reconstruction Game: Drawing Common Ground Through Iterative Multimodal Dialogue

    Sherzod Hakimov, Mattia D'Agostini, Ivan Samodelkin, David Schlangen

    cs.CV · cs.AI · cs.CL

    We introduce the Image Reconstruction Game, a fully automated benchmark in which a vision-language model issues corrective instructions to an image generator across multiple turns, making accumulated common ground directly observable as a rendered image. Benchmarking two Describer models crossed with two Generator models across seven image categories, we find that the describer is the dominant factor in reconstruction quality, while the...

    arxiv.org/abs/2606.01901 · PDF

  28. 28

    Train, Test, Re-evaluate: Schedule-Sensitive Evaluation of Generative Data for Hand Detection

    Atmika Bhardwaj, Silvia Vock, Nico Steckhan

    cs.CV · cs.AI

    Generated (or synthetic) image data is increasingly used to augment or replace real training datasets when target imagery is scarce, expensive, or biased. For hand detection, particularly in occupational safety settings, public datasets mostly contain bare hands. This under-represents the variation in hand appearance introduced by gloves, tattoos, jewelry, and other personal protective equipment, creating a distribution shift that...

    arxiv.org/abs/2606.01896 · PDF

  29. 29

    Collaborative Space Object Detection with Multi-Satellite Viewpoints in LEO Constellations

    Xingyu Qu, Wenxuan Zhang, Peng Hu

    cs.CV · cs.AI

    With the growing number of satellites in low Earth orbit (LEO) constellations, the near-Earth space environment has become increasingly congested, making space object detection (SOD) a pressing challenge for space safety and sustainability. To mitigate collision risks and ensure the continuity of space operations, SOD systems must deliver fast and accurate detection under stringent onboard constraints. In this paper, we investigate the...

    arxiv.org/abs/2606.01895 · PDF

This edition is part of The Daily Abstract — cs.CV archive. Subscribe to receive these in your inbox each morning, automatically translated to Spanish, with reply-to-PDF: arxivdaily.ignorelist.com.

Colophon Set in Georgia, with system sans for interface chrome and a monospaced stack for code and paper identifiers. Sole accent: amber #D99C5E. Built and served on an always-free VM. The masthead is set 14% letterspaced because newspapers do that and it works.