视觉大模型 / VLM - arXivDaily 专题

2505.18675 2026-03-13 cs.CV cs.AI cs.CL 86%

ReasonMap: Towards Fine-Grained Visual Reasoning from Transit Maps

ReasonMap：迈向基于交通地图的细粒度视觉推理

Sicheng Feng, Song Wang, Shuyi Ouyang, Lingdong Kong, Zikai Song, Jianke Zhu, Huan Wang, Xinchao Wang

专题命中视觉推理：visual reasoning(title,abstract)；grounding(abstract)；multimodal large language model(abstract)；分类 cs.CV、cs.AI

AI总结 ReasonMap是一个针对交通地图的细粒度视觉推理基准，通过评估16种MLLMs揭示开源与闭源模型在推理性能上的差异，并强调视觉支撑对高性能的重要性。

Comments CVPR 2026, website: https://fscdc.github.io/ReasonMap/

2602.14482 2026-03-12 cs.CV cs.AI 86%

TikArt: Stabilizing Aperture-Guided Fine-Grained Visual Reasoning with Reinforcement Learning

TikArt: 通过强化学习稳定引导孔细粒度视觉推理

Hao Ding, Zhichuan Yang, Weijie Ge, Ziqin Gao, Chaoyi Lu, Lei Zhao

专题命中视觉推理：visual reasoning(title,abstract)；grounding(abstract)；multimodal large language model(abstract)；分类 cs.CV、cs.AI

AI总结 TikArt通过强化学习引导孔机制提升细粒度视觉推理，结合语言推理与Zoom、Segment操作，实现多模态证据获取与持久记忆。

2510.20812 2026-03-02 cs.CV cs.AI cs.CL 86%

Small Drafts, Big Verdict: Information-Intensive Visual Reasoning via Speculation

小草稿，大结论：通过推测进行信息密集的视觉推理

Yuhan Liu, Lianhui Qin, Shengjie Wang

机构 * New York University（纽约大学）； University of California, San Diego（加州大学圣地亚哥分校）

专题命中视觉推理：visual reasoning(title)；vision-language model(abstract)；VLM(abstract)；visual question answering(abstract)

AI总结 SV通过结合多个轻量级草稿专家和大型结论模型，提升信息密集视觉推理的效率和准确性。

Comments Accepted to ICLR 2026

2510.02240 2026-02-24 cs.CV cs.AI 86%

RewardMap: Tackling Sparse Rewards in Fine-grained Visual Reasoning via Multi-Stage Reinforcement Learning

RewardMap: 通过多阶段强化学习解决细粒度视觉推理中的稀疏奖励

Sicheng Feng, Kaiwen Tuo, Song Wang, Lingdong Kong, Jianke Zhu, Huan Wang

机构 * Westlake University（西湖大学）； Tongji University（同济大学）； Zhejiang University（浙江大学）； National University of Singapore（新加坡国立大学）

专题命中视觉推理：visual reasoning(title,abstract)；visual question answering(abstract)；multimodal large language model(abstract)；分类 cs.CV、cs.AI

AI总结 RewardMap通过多阶段强化学习解决细粒度视觉推理中的稀疏奖励问题，提升多模态大语言模型的视觉理解和推理能力。

Comments ICLR 2026, website: https://fscdc.github.io/RewardMap/

2601.02422 2026-01-07 cs.CV cs.AI 86%

Watch Wider and Think Deeper: Collaborative Cross-modal Chain-of-Thought for Complex Visual Reasoning

拓宽视野并深入思考：协作式跨模态链式推理用于复杂视觉推理

Wenting Lu, Didi Zhu, Tao Shen, Donglin Zhu, Ayong Ye, Chao Wu

机构 * Fujian Normal University（福建师范大学）； Zhejiang University（浙江大学）； Zhejiang Normal University（浙江师范大学）

专题命中视觉推理：visual reasoning(title,abstract)；LLaVA(abstract)；grounding(abstract)；分类 cs.CV、cs.AI

AI总结 CoCoT通过动态多区域定位和关系感知推理，提升复杂视觉推理性能，在多个基准测试中实现显著准确率提升。

2512.22183 2025-12-30 cs.CV cs.AI cs.CL 86%

Unbiased Visual Reasoning with Controlled Visual Inputs

通过受控视觉输入实现无偏视觉推理

Zhaonan Li, Shijie Lu, Fei Wang, Jacob Dineen, Xiao Ye, Zhikun Xu, Siyi Liu, Young Min Cho, Bangzheng Li, Daniel Chang, Kenny Nguyen, Qizheng Yang, Muhao Chen, Ben Zhou

机构 * Arizona State University（亚利桑那州立大学）； University of Southern California（南加州大学）； University of Pennsylvania（宾夕法尼亚大学）； University of California, Davis（加州大学戴维斯分校）

专题命中视觉推理：visual reasoning(title,abstract)；vision-language model(abstract)；VLM(abstract)；分类 cs.CV、cs.AI

AI总结 VISTA通过解耦感知与推理，利用受控接口和强化学习训练，提升了视觉推理的鲁棒性和中立性。

2508.06259 2025-12-29 cs.CV cs.AI 86%

SIFThinker: Spatially-Aware Image Focus for Visual Reasoning

SIFThinker: 基于空间的图像聚焦用于视觉推理

Zhangquan Chen, Ruihui Zhao, Chuwei Luo, Mingze Sun, Xinlei Yu, Yangyang Kang, Ruqi Huang

机构 * ByteDance（字节跳动）

专题命中视觉推理：visual reasoning(title,abstract)；grounding(abstract)；multimodal large language model(abstract)；分类 cs.CV、cs.AI

AI总结 SIFThinker通过结合深度增强的边界框和自然语言，提升视觉推理中的空间理解和细粒度感知能力。

Comments 15 pages, 13 figures

2512.15885 2025-12-19 cs.CV cs.AI cs.CL cs.MM 86%

Seeing Beyond Words: Self-Supervised Visual Learning for Multimodal Large Language Models

超越文字：面向多模态大语言模型的自监督视觉学习

Davide Caffagni, Sara Sarto, Marcella Cornia, Lorenzo Baraldi, Pier Luigi Dovesi, Shaghayegh Roohi, Mark Granroth-Wilding, Rita Cucchiara

机构 * University of Modena and Reggio Emilia（摩德纳和雷吉奥艾米利亚大学）； AMD Silo AI

专题命中视觉推理：multimodal large language model(title,abstract)；visual reasoning(abstract)；MLLM(abstract)；分类 cs.CV、cs.AI

AI总结 JARVIS通过自监督视觉学习提升多模态大语言模型的视觉推理能力，无需依赖语言监督。

2510.06067 2025-11-18 cs.CV cs.AI 86%

Reasoning under Vision: Understanding Visual-Spatial Cognition in Vision-Language Models for CAPTCHA

Python Song, Luke Tenyi Chang, Yun-Yun Tsai, Penghui Li, Junfeng Yang

机构 * Columbia University Department of Computer Science（哥伦比亚大学计算机科学系）

专题命中视觉推理：vision-language model(title,abstract)；VLM(abstract)；grounding(abstract)；分类 cs.CV、cs.AI

Comments 14pages, 11figures

2510.24152 2025-10-29 cs.CV cs.AI 86%

Enhancing Vision-Language Models for Autonomous Driving through Task-Specific Prompting and Spatial Reasoning

Aodi Wu, Xubo Luo

机构 * University of Chinese Academy of Sciences（中国科学院大学）； Technology and Engineering Center for Space Utilization, Chinese Academy of Sciences（中国科学院空间利用技术与工程中心）

专题命中视觉推理：vision-language model(title,abstract)；VLM(abstract)；grounding(abstract)；分类 cs.CV、cs.AI

Comments RoboSense Challenge with IROS 2025

2510.20967 2025-10-27 cs.CV cs.AI 86%

3DReasonKnee: Advancing Grounded Reasoning in Medical Vision Language Models

Sraavya Sambara, Sung Eun Kim, Xiaoman Zhang, Luyang Luo, Shreya Johri, Mohammed Baharoon, Du Hyun Ro, Pranav Rajpurkar

机构 * Department of Biomedical Informatics, Harvard Medical School（生物医学信息学系，哈佛医学院）； Seoul National University Hospital（首尔国立大学医院）

专题命中视觉推理：vision language model(title)；vision-language model(abstract)；VLM(abstract)；grounding(abstract)

2506.16962 2025-10-23 cs.CV cs.AI cs.CL 86%

Chiron-o1: Igniting Multimodal Large Language Models towards Generalizable Medical Reasoning via Mentor-Intern Collaborative Search

Haoran Sun, Yankai Jiang, Wenjie Lou, Yujie Zhang, Wenjie Li, Lilong Wang, Mianxin Liu, Lei Liu, Xiaosong Wang

机构 * Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）； Fudan University（复旦大学）； Shanghai Jiao Tong University（上海交通大学）

专题命中视觉推理：multimodal large language model(title,abstract)；visual question answering(abstract)；MLLM(abstract)；分类 cs.CV、cs.AI

2510.12845 2025-10-16 cs.CL cs.AI cs.CV cs.RO 86%

VLURes: Benchmarking VLM Visual and Linguistic Understanding in Low-Resource Languages

Jesse Atuhurra, Iqra Ali, Tomoya Iwakura, Hidetaka Kamigaito, Tatsuya Hiraoka

专题命中视觉推理：VLM(title,abstract)；vision language model(abstract)；visual reasoning(abstract)；分类 cs.CV、cs.AI

2510.11718 2025-10-14 cs.CV cs.AI 86%

CodePlot-CoT: Mathematical Visual Reasoning by Thinking with Code-Driven Images

Chengqi Duan, Kaiyue Sun, Rongyao Fang, Manyuan Zhang, Yan Feng, Ying Luo, Yufang Liu, Ke Wang, Peng Pei, Xunliang Cai, Hongsheng Li, Yi Ma, Xihui Liu

机构 * HKU（香港大学）； Meituan（美团）； CUHK（香港中文大学）

专题命中视觉推理：visual reasoning(title,abstract)；vision language model(abstract)；VLM(abstract)；分类 cs.CV、cs.AI

2508.02095 2025-08-08 cs.CV cs.AI 86%

VLM4D: Towards Spatiotemporal Awareness in Vision Language Models

Shijie Zhou, Alexander Vilesov, Xuehai He, Ziyu Wan, Shuwang Zhang, Aditya Nagachandra, Di Chang, Dongdong Chen, Xin Eric Wang, Achuta Kadambi

机构 * UCLA（美国大学洛杉矶分校）； Microsoft（微软公司）； UCSC（加州大学圣塔克拉拉分校）； USC（美国大学洛杉矶分校）

专题命中视觉推理：vision language model(title,abstract)；visual reasoning(abstract)；grounding(abstract)；分类 cs.CV、cs.AI

Comments ICCV 2025, Project Website: https://vlm4d.github.io/

2507.01029 2025-07-03 cs.LG cs.AI cs.CL 86%

PathCoT: Chain-of-Thought Prompting for Zero-shot Pathology Visual Reasoning

Junjie Zhou, Yingli Zuo, Shichang Feng, Peng Wan, Qi Zhu, Daoqiang Zhang, Wei Shao

机构 * The College of Artificial Intelligence, Nanjing University of Aeronautics（人工智能学院，南京航空航天大学）； Astronautics The Key Laboratory of Brain-Machine Intelligence Technology, Ministry of Education（航空航天脑机智能技术重点实验室，教育部）

专题命中视觉推理：visual reasoning(title,abstract)；multimodal large language model(abstract)；MLLM(abstract)；分类 cs.AI、cs.LG

2506.19469 2025-06-25 cs.CV cs.AI 86%

Surgery-R1: Advancing Surgical-VQLA with Reasoning Multimodal Large Language Model via Reinforcement Learning

Pengfei Hao, Shuaibo Li, Hongqiu Wang, Zhizhuo Kou, Junhang Zhang, Guang Yang, Lei Zhu

机构 * Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））； Hong Kong University of Science and Technology（香港科技大学）； Department of Thoracic Surgery, the Seventh Affiliated Hospital, Sun Yat-sen University（中山大学第七附属医院胸外科部门）； Bioengineering/Imperial-X, Imperial College London（生物工程/Imperial-X，帝国理工学院伦敦分校）； ROAS Thrust, Hong Kong University of Science and Technology (Guangzhou)（ROAS项目，香港科技大学（广州））； Department of Electronic and Computer Engineering, Hong Kong SAR（香港特别行政区电子与计算机工程系）

专题命中视觉推理：multimodal large language model(title,abstract)；grounding(abstract)；MLLM(abstract)；分类 cs.CV、cs.AI

2501.02669 2025-06-03 cs.CV cs.CL cs.LG 86%

Generalizing from SIMPLE to HARD Visual Reasoning: Can We Mitigate Modality Imbalance in VLMs?

Simon Park, Abhishek Panigrahi, Yun Cheng, Dingli Yu, Anirudh Goyal, Sanjeev Arora

专题命中视觉推理：visual reasoning(title,abstract)；vision language model(abstract)；visual question answering(abstract)；分类 cs.CV、cs.LG

2505.16192 2025-06-02 cs.CV cs.AI 86%

VLM-R$^3$: Region Recognition, Reasoning, and Refinement for Enhanced Multimodal Chain-of-Thought

Chaoya Jiang, Yongrui Heng, Wei Ye, Han Yang, Haiyang Xu, Ming Yan, Ji Zhang, Fei Huang, Shikun Zhang

机构 * National Engineering Research Center for Software Engineering, Peking University（软件工程国家工程研究中心，北京大学）； Alibaba Group（阿里巴巴集团）； ZEEKR Intelligent Technology Holding Limited（ZEKR智能科技控股有限公司）

专题命中视觉推理：VLM(title,abstract)；grounding(abstract)；MLLM(abstract)；分类 cs.CV、cs.AI

2504.20648 2025-04-30 cs.CV cs.AI 86%

SpaRE: Enhancing Spatial Reasoning in Vision-Language Models with Synthetic Data

Michael Ogezi, Freda Shi

机构 * University of Waterloo（滑铁卢大学）； Vector Institute（向量研究所）

专题命中视觉推理：vision-language model(title,abstract)；VLM(abstract)；visual question answering(abstract)；分类 cs.CV、cs.AI

2504.02587 2025-04-07 cs.LG cs.CL cs.CV 86%

Rethinking RL Scaling for Vision Language Models: A Transparent, From-Scratch Framework and Comprehensive Evaluation Scheme

Yan Ma, Steffi Chern, Xuyang Shen, Yiran Zhong, Pengfei Liu

专题命中视觉推理：vision language model(title)；vision-language model(abstract)；VLM(abstract)；visual reasoning(abstract)

Comments Code is public and available at: https://github.com/GAIR-NLP/MAYE

2502.19634 2025-03-20 cs.CV cs.AI 86%

MedVLM-R1: Incentivizing Medical Reasoning Capability of Vision-Language Models (VLMs) via Reinforcement Learning

Jiazhen Pan, Che Liu, Junde Wu, Fenglin Liu, Jiayuan Zhu, Hongwei Bran Li, Chen Chen, Cheng Ouyang, Daniel Rueckert

专题命中视觉推理：vision-language model(title)；VLM(abstract)；visual language model(abstract)；visual question answering(abstract)

2501.18880 2025-02-03 cs.CV cs.LG 86%

RLS3: RL-Based Synthetic Sample Selection to Enhance Spatial Reasoning in Vision-Language Models for Indoor Autonomous Perception

Joshua R. Waite, Md. Zahid Hasan, Qisai Liu, Zhanhong Jiang, Chinmay Hegde, Soumik Sarkar

专题命中视觉推理：vision-language model(title,abstract)；VLM(abstract)；grounding(abstract)；分类 cs.CV、cs.LG

Comments ICCPS 2025 accepted paper, 10 pages, 9 figures

2407.16837 2025-01-14 cs.CV cs.AI cs.CL 86%

MLLM-CompBench: A Comparative Reasoning Benchmark for Multimodal LLMs

Jihyung Kil, Zheda Mai, Justin Lee, Zihe Wang, Kerrie Cheng, Lemeng Wang, Ye Liu, Arpita Chowdhury, Wei-Lun Chao

专题命中视觉推理：MLLM(title,abstract)；LLaVA(abstract)；multimodal large language model(abstract)；分类 cs.CV、cs.AI

Comments This paper has been accepted to NeurIPS 2024. The first two authors contributed equally to this work

2412.08125 2024-12-20 cs.CV cs.CL cs.LG 86%

Progressive Multi-granular Alignments for Grounded Reasoning in Large Vision-Language Models

Quang-Hung Le, Long Hoang Dang, Ngan Le, Truyen Tran, Thao Minh Le

专题命中视觉推理：vision-language model(title,abstract)；visual reasoning(abstract)；grounding(abstract)；分类 cs.CV、cs.LG

2409.13980 2024-09-24 cs.CV cs.AI 86%

Enhancing Advanced Visual Reasoning Ability of Large Language Models

Zhiyuan Li, Dongnan Liu, Chaoyi Zhang, Heng Wang, Tengfei Xue, Weidong Cai

专题命中视觉推理：visual reasoning(title,abstract)；vision-language model(abstract)；multimodal large language model(abstract)；分类 cs.CV、cs.AI

Comments EMNLP 2024 Main

2401.12168 2024-01-23 cs.CV cs.CL cs.LG cs.RO 86%

SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities

Boyuan Chen, Zhuo Xu, Sean Kirmani, Brian Ichter, Danny Driess, Pete Florence, Dorsa Sadigh, Leonidas Guibas, Fei Xia

专题命中视觉推理：vision-language model(title)；vision language model(abstract)；VLM(abstract)；visual question answering(abstract)

2605.20177 2026-05-20 cs.CL cs.CV 86%

From Seeing to Thinking: Decoupling Perception and Reasoning Improves Post-Training of Vision-Language Models

从感知到思考：解耦感知与推理提升视觉语言模型的训练

Juncheng Wu, Hardy Chen, Haoqin Tu, Xianfeng Tang, Freda Shi, Hui Liu, Hanqing Lu, Cihang Xie, Yuyin Zhou

机构 * Amazon（亚马逊）； University of Waterloo（滑铁卢大学）； Vector Institute, Canada CIFAR AI Chair（向量研究所，加拿大CIFAR人工智能主席）

专题命中视觉推理：vision-language model(title,abstract)；VLM(abstract,comments)；visual reasoning(abstract)；分类 cs.CV

AI总结本研究通过解耦感知与推理，发现视觉任务性能受限于感知能力不足而非推理本身，通过分阶段训练提升模型的感知与推理能力，从而在多个视觉数学和感知任务中取得更优表现。

Comments 19 pages, 9 figures; Accepted to ICML 2026; Project Page: https://ucsc-vlaa.github.io/VLM-CapCurriculum/

2511.19418 2026-08-06 cs.CV cs.AI cs.LG 版本更新 86%

Chain-of-Visual-Thought: Teaching VLMs to See and Think Better with Continuous Visual Tokens

链式视觉思维：通过连续视觉标记教学VLMs更好地看到和思考

Yiming Qin, Bomin Wei, Jiaxin Ge, Konstantinos Kallidromitis, Stephanie Fu, Trevor Darrell, XuDong Wang

机构 * UC Berkeley（加州大学伯克利分校）； UCLA（洛杉矶大学）； Panasonic AI Research（松下人工智能研究）

专题命中视觉推理：VLM(abstract,abstract_cn)；LLaVA(abstract,abstract_cn)；vision-language model(abstract)；分类 cs.CV、cs.AI、cs.LG

AI总结 COVT通过连续视觉标记提升VLMs的视觉推理能力，使模型在多个基准测试中性能提升3%-16%。

Comments Project page: https://wakalsprojectpage.github.io/covt-website/

2511.01472 2026-06-30 cs.RO 86%

AERMANI-VLM: Structured Prompting and Reasoning for Aerial Manipulation with Vision Language Models

AERMANI-VLM：基于视觉语言模型的空中 manipulation 的结构提示与推理

Sarthak Mishra, Rishabh Dev Yadav, Avirup Das, Saksham Gupta, Wei Pan, Spandan Roy

机构 * Robotics Research Center, IIIT Hyderabad（IIIT海得拉巴机器人研究中心）； Department of Computer Science, University of Manchester（曼彻斯特大学计算机科学系）； Newcastle University（纽卡斯尔大学）

专题命中视觉推理：VLM(title,abstract)；vision language model(title)

AI总结本文提出AERMANI-VLM框架，通过分离高层推理与低层控制，利用结构化提示生成自然语言推理轨迹，实现安全可靠的空中 manipulation 任务执行。

1. 视觉推理 4478 篇

ReasonMap: Towards Fine-Grained Visual Reasoning from Transit Maps

TikArt: Stabilizing Aperture-Guided Fine-Grained Visual Reasoning with Reinforcement Learning

Small Drafts, Big Verdict: Information-Intensive Visual Reasoning via Speculation

RewardMap: Tackling Sparse Rewards in Fine-grained Visual Reasoning via Multi-Stage Reinforcement Learning

Watch Wider and Think Deeper: Collaborative Cross-modal Chain-of-Thought for Complex Visual Reasoning

Unbiased Visual Reasoning with Controlled Visual Inputs

SIFThinker: Spatially-Aware Image Focus for Visual Reasoning

Seeing Beyond Words: Self-Supervised Visual Learning for Multimodal Large Language Models

Reasoning under Vision: Understanding Visual-Spatial Cognition in Vision-Language Models for CAPTCHA

Enhancing Vision-Language Models for Autonomous Driving through Task-Specific Prompting and Spatial Reasoning

3DReasonKnee: Advancing Grounded Reasoning in Medical Vision Language Models

Chiron-o1: Igniting Multimodal Large Language Models towards Generalizable Medical Reasoning via Mentor-Intern Collaborative Search

VLURes: Benchmarking VLM Visual and Linguistic Understanding in Low-Resource Languages

CodePlot-CoT: Mathematical Visual Reasoning by Thinking with Code-Driven Images

VLM4D: Towards Spatiotemporal Awareness in Vision Language Models

PathCoT: Chain-of-Thought Prompting for Zero-shot Pathology Visual Reasoning

Surgery-R1: Advancing Surgical-VQLA with Reasoning Multimodal Large Language Model via Reinforcement Learning

Generalizing from SIMPLE to HARD Visual Reasoning: Can We Mitigate Modality Imbalance in VLMs?

VLM-R$^3$: Region Recognition, Reasoning, and Refinement for Enhanced Multimodal Chain-of-Thought

SpaRE: Enhancing Spatial Reasoning in Vision-Language Models with Synthetic Data

Rethinking RL Scaling for Vision Language Models: A Transparent, From-Scratch Framework and Comprehensive Evaluation Scheme

MedVLM-R1: Incentivizing Medical Reasoning Capability of Vision-Language Models (VLMs) via Reinforcement Learning

RLS3: RL-Based Synthetic Sample Selection to Enhance Spatial Reasoning in Vision-Language Models for Indoor Autonomous Perception

MLLM-CompBench: A Comparative Reasoning Benchmark for Multimodal LLMs

Progressive Multi-granular Alignments for Grounded Reasoning in Large Vision-Language Models

Enhancing Advanced Visual Reasoning Ability of Large Language Models

SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities

From Seeing to Thinking: Decoupling Perception and Reasoning Improves Post-Training of Vision-Language Models

Chain-of-Visual-Thought: Teaching VLMs to See and Think Better with Continuous Visual Tokens

AERMANI-VLM: Structured Prompting and Reasoning for Aerial Manipulation with Vision Language Models