CoChat
Sign Up
Cited by 6Multimodal Machine Learning ApplicationsAdvanced Neural Network ApplicationsVisual Attention and Saliency Detection

SegAgent: Exploring Pixel Understanding Capabilities in MLLMs by Imitating Human Annotator Trajectories

Open access
Zhu, Muzhi, Tian, Yuzhuo, Chen, Hao, Zhou, Chunluan, Guo, Qingpei, Liu, Yang, Yang, Ming, Shen, Chunhua·2025·2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
DOI 10.1109/cvpr52734.2025.00349
Open this paper in CoChat — free accountOpen PDF

Abstract

While MLLMs have demonstrated impressive image understanding capabilities, they still struggle with pixel-level comprehension, limiting their practical applications. Current evaluation tasks such as VQA and visual grounding remain too coarse to assess fine-grained pixel comprehension accurately. Although segmentation is foundational for pixel-level understanding, existing methods often require MLLMs to generate implicit tokens, decoded through external pixel decoders. This approach disrupts the MLLM’s text output space, potentially compromising language capabilities and reducing flexibility and extensibility while failing to reflect the model’s intrinsic pixel-level understanding. Thus, we introduce the Human-Like Mask Annotation Task (HLMAT), a new paradigm where MLLMs mimic human annotators using interactive segmentation tools. Modelling segmentation as a multi-step Markov Decision Process, HLMAT enables MLLMs to iteratively generate text-based click points, achieving high-quality masks without architectural changes or implicit tokens. Through this setup, we develop SegAgent, a model fine-tuned on human-like annotation trajectories, which achieves performance comparable to SoTA methods and supports additional tasks like mask refinement and annotation filtering. HLMAT provides a protocol for assessing fine-grained pixel understanding in MLLMs and introduces a vision-centric, multi-step decision-making task that facilitates the exploration of MLLMs’ visual reasoning abilities. Our adaptations of policy improvement method StaR and PRM guided tree search further enhance model robustness in complex segmentation tasks, laying a foundation for future advancements in fine-grained visual perception and multi-step decision-making for MLLMs. Code can be found at https://github.com/aim-uofa/SegAgent.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/cvpr52734.2025.00349

Cites

34 of 34 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Existing methods typically restrict the action space to sequential clicks.

    “they typically restrict the action space to sequential clicks.”

    From MedSAM-Agent: Empowering Interactive Medical Image Segmentation with Multi-turn Agentic Reinforcement Learning (2026) · cites this paper as Zhu et al., 2025

    Section 4.1 (p. 4) describes an existing iterative click-simulation strategy, and Section 5.1 (p. 6) says SimpleClick accepts only points. But Section 2.3 (p. 3) also describes interactive segmentation using boxes and scribbles, so the paper does not establish that existing methods typically restrict their action space to sequential clicks. (LLM, conf=0.76)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 42

  1. [1][1] Hao Bai, Yifei Zhou, Mert Cemri, Jiayi Pan, Alane Suhr, Sergey Levine, and Aviral Kumar. Digirl: Training in-thewild device-control agents with autonomous reinforcement learning. arXiv preprint arXiv:2406.11896 , 2024. - · arXiv
  2. [2][2] Jinze Bai, Shuai Bai, Shusheng Yang, Shijie Wang, Sinan Tan, Peng Wang, Junyang Lin, Chang Zhou, and Jingren Zhou. Qwen-vl: A frontier large vision-language model with versatile abilities. arXiv preprint arXiv:2308.12966 , 2023. - · arXiv
  3. [4][4] Shengcao Cao, Liang-Yan Gui, and Yu-Xiong Wang. Emerging pixel grounding in large multimodal models without grounding supervision. arXiv preprint arXiv:2410.08209 , 2024. - · arXiv
  4. [7][7] Ting Chen, Saurabh Saxena, Lala Li, Tsung-Yi Lin, David J Fleet, and Geoffrey E Hinton. A unified sequence interface for vision tasks. Advances in Neural Information Processing Systems , 35:31333-31346, 2022. -
  5. [8][8] William Chen, Oier Mees, Aviral Kumar, and Sergey Levine. Vision-language models provide promptable representations for reinforcement learning. arXiv preprint arXiv:2402.02651 , 2024. - · arXiv
  6. [13][13] Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948 , 2025. - · arXiv
  7. [16][16] Lei Ke, Mingqiao Ye, Martin Danelljan, Yu-Wing Tai, ChiKeung Tang, Fisher Yu, et al. Segment anything in high quality. Advances in Neural Information Processing Systems , 36, 2024. -
  8. [19][19] Xin Lai, Zhuotao Tian, Yukang Chen, Yanwei Li, Yuhui Yuan, Shu Liu, and Jiaya Jia. Lisa: Reasoning segmentation via large language model. arXiv preprint arXiv:2308.00692 , 2023. - · arXiv
  9. [22][22] Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation. In International Conference on Machine Learning , pages 1288812900. PMLR, 2022. -
  10. [23][23] Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. arXiv preprint arXiv:2301.12597 , 2023. - · arXiv
  11. [25][25] Hunter Lightman, Vineet Kosaraju, Yura Burda, Harri Edwards, Bowen Baker, Teddy Lee, Jan Leike, John Schulman, Ilya Sutskever, and Karl Cobbe. Let's verify step by step. arXiv preprint arXiv:2305.20050 , 2023. - · arXiv
  12. [29][29] Jiang Liu, Hui Ding, Zhaowei Cai, Yuting Zhang, Ravi Kumar Satzoda, Vijay Mahadevan, and R Manmatha. Polyformer: Referring image segmentation as sequential polygon generation. In CVPR , 2023. -
  13. [31][31] Yang Liu, Muzhi Zhu, Hengtao Li, Hao Chen, Xinlong Wang, and Chunhua Shen. Matcher: Segment anything with one shot using all-purpose feature matching. arXiv preprint arXiv:2305.13310 , 2023. - · arXiv
  14. [32][32] Yang Liu, Chenchen Jing, Hengtao Li, Muzhi Zhu, Hao Chen, Xinlong Wang, and Chunhua Shen. A simple image segmentation framework via in-context examples. Advances in Neural Information Processing Systems , 37:25095-25119, 2025. -
  15. [33][33] Yuqi Liu, Bohao Peng, Zhisheng Zhong, Zihao Yue, Fanbin Lu, Bei Yu, and Jiaya Jia. Seg-zero: Reasoning-chain guided segmentation via cognitive reinforcement. arXiv preprint arXiv:2503.06520 , 2025. - · arXiv
  16. [34][34] Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, and Saining Xie. A convnet for the 2020s. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 11976-11986, 2022. -
  17. [35][35] Ilya Loshchilov and Frank Hutter. Sgdr: Stochastic gradient descent with warm restarts. arXiv preprint arXiv:1608.03983 , 2016. - · arXiv
  18. [36][36] Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 , 2017. - · arXiv
  19. [37][37] Qianli Ma, Haotian Zhou, Tingkai Liu, Jianbo Yuan, Pengfei Liu, Yang You, and Hongxia Yang. Let's reward step by step: Step-level reward model as the navigators for reasoning. arXiv preprint arXiv:2310.10080 , 2023. - · arXiv
  20. [39][39] Yao Mu, Qinglong Zhang, Mengkang Hu, Wenhai Wang, Mingyu Ding, Jun Jin, Bin Wang, Jifeng Dai, Yu Qiao, and Ping Luo. Embodiedgpt: Vision-language pretraining via embodied chain of thought. arXiv preprint arXiv:2305.15021 , 2023. - · arXiv
  21. [41][41] Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems , 35: 27730-27744, 2022. -
  22. [48][48] Mennatullah Siam. Pixfoundation: Are we heading in the right direction with pixel-level vision foundation models? arXiv preprint arXiv:2502.04192 , 2025. - · arXiv
  23. [50][50] Gemini Team, Rohan Anil, Sebastian Borgeaud, Yonghui Wu, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, et al. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805 , 2023. - · arXiv
  24. [51][51] Shengbang Tong, Ellis Brown, Penghao Wu, Sanghyun Woo, Manoj Middepogu, Sai Charitha Akula, Jihan Yang, Shusheng Yang, Adithya Iyer, Xichen Pan, et al. Cambrian1: A fully open, vision-centric exploration of multimodal llms. arXiv preprint arXiv:2406.16860 , 2024. - · arXiv
  25. [53][53] Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, et al. Qwen2-vl: Enhancing vision-language model's perception of the world at any resolution. arXiv preprint arXiv:2409.12191 , 2024. - · arXiv
  26. [55][55] Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems , 35:24824-24837, 2022. -
  27. [57][57] Senqiao Yang, Tianyuan Qu, Xin Lai, Zhuotao Tian, Bohao Peng, Shu Liu, and Jiaya Jia. An improved baseline for reasoning segmentation with large language model. arXiv preprint arXiv:2312.17240 , 2023. - · arXiv
  28. [60][60] Linli Yao, Lei Li, Shuhuai Ren, Lean Wang, Yuanxin Liu, Xu Sun, and Lu Hou. Deco: Decoupling token compression from semantic abstraction in multimodal large language models. arXiv preprint arXiv:2405.20985 , 2024. - · arXiv
  29. [61][61] Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Thomas L Griffiths, Yuan Cao, and Karthik Narasimhan. 15. Tree of thoughts: Deliberate problem solving with large language models. arXiv preprint arXiv:2305.10601 , 2023. - · arXiv
  30. [62][62] Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R Narasimhan, and Yuan Cao. React: Synergizing reasoning and acting in language models. In The Eleventh International Conference on Learning Representations , 2023. -
  31. [63][63] Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Tom Griffiths, Yuan Cao, and Karthik Narasimhan. Tree of thoughts: Deliberate problem solving with large language models. Advances in Neural Information Processing Systems , 36, 2024. -
  32. [64][64] Kaining Ying, Fanqing Meng, Jin Wang, Zhiqian Li, Han Lin, Yue Yang, Hao Zhang, Wenbo Zhang, Yuqi Lin, Shuo Liu, et al. Mmt-bench: A comprehensive multimodal benchmark for evaluating large vision-language models towards multitask agi. arXiv preprint arXiv:2404.16006 , 2024. - · arXiv
  33. [66][66] Licheng Yu, Patrick Poirson, Shan Yang, Alexander C Berg, and Tamara L Berg. Modeling context in referring expressions. In ECCV , 2016. -
  34. [68][68] Hongyi Yuan, Zheng Yuan, Chuanqi Tan, Wei Wang, Songfang Huang, and Fei Huang. Rrhf: Rank responses to align language models with human feedback. Advances in Neural Information Processing Systems , 36, 2024. -
  35. [70][70] Eric Zelikman, Yuhuai Wu, Jesse Mu, and Noah Goodman. Star: Bootstrapping reasoning with reasoning. Advances in Neural Information Processing Systems , 35:15476-15488, 2022. -
  36. [71][71] Yuexiang Zhai, Hao Bai, Zipeng Lin, Jiayi Pan, Shengbang Tong, Yifei Zhou, Alane Suhr, Saining Xie, Yann LeCun, Yi Ma, et al. Fine-tuning large vision-language models as decision-making agents via reinforcement learning. arXiv preprint arXiv:2405.10292 , 2024. - · arXiv
  37. [73][73] Tao Zhang, Xiangtai Li, Hao Fei, Haobo Yuan, Shengqiong Wu, Shunping Ji, Chen Change Loy, and Shuicheng Yan. Omg-llava: Bridging image-level, object-level, pixel-level reasoning and understanding. arXiv preprint arXiv:2406.19389 , 2024. - · arXiv
  38. [76][76] Canyu Zhao, Mingyu Liu, Huanyi Zheng, Muzhi Zhu, Zhiyue Zhao, Hao Chen, Tong He, and Chunhua Shen. Diception: A generalist diffusion model for visual perceptual tasks. arXiv preprint arXiv:2502.17157 , 2025. - · arXiv
  39. [77][77] Banghua Zhu, Hiteshi Sharma, Felipe Vieira Frujeri, Shi Dong, Chenguang Zhu, Michael I Jordan, and Jiantao Jiao. Fine-tuning language models with advantage-induced policy alignment. arXiv preprint arXiv:2306.02231 , 2023. - · arXiv
  40. [79][79] Muzhi Zhu, Chengxiang Fan, Hao Chen, Yang Liu, Weian Mao, Xiaogang Xu, and Chunhua Shen. Generative active learning for long-tailed instance segmentation. arXiv preprint arXiv:2406.02435 , 2024. - · arXiv
  41. [80][80] Muzhi Zhu, Yang Liu, Zekai Luo, Chenchen Jing, Hao Chen, Guangkai Xu, Xinlong Wang, and Chunhua Shen. Unleashing the potential of the diffusion model in few-shot semantic segmentation. arXiv preprint arXiv:2410.02369 , 2024. - · arXiv
  42. [82][82] Xueyan Zou, Jianwei Yang, Hao Zhang, Feng Li, Linjie Li, Jianfeng Wang, Lijuan Wang, Jianfeng Gao, and Yong Jae Lee. Segment everything everywhere all at once. NeurIPS , 2024.

Cited by

Sorted by citations

Related

Authors
Zhu, M., Tian, Y., Chen, H., Zhou, C., Guo, Q., Liu, Y., Yang, M., & Shen, C.
Affiliations
  • Zhejiang University (CN)
  • Ant Group (China) (CN)
ORCID
Venue
2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Publisher
IEEE
Published
2025-06-10
Pages
3686-3696
ISSN
1063-6919 (ISSN-L)
DOI
10.1109/cvpr52734.2025.00349
OpenAlex
W4413147628
Cited by
5
References
34
Citations by year
2026: 5
FWCI
3.78
Type
conference paper
Language
en
Open access
Licence
Locations
  • IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • National Natural Science Foundation of China
  • Ant Group
Topics
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Visual Attention and Saliency Detection
Keywords
multimodal large language models, image segmentation, interactive segmentation, Markov decision process, visual grounding, visual reasoning, mask refinement, stars

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review