MedReasoner: Reinforcement Learning Drives Reasoning Grounding from Clinical Thought to Pixel-Level Precision
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918343438499840 |
|---|---|
| author | Yan, Zhonghao Diao, Muxi Yang, Yuxuan Jing, Ruoyan Xu, Jiayuan Zhang, Kaizhou Yang, Lele Liu, Yanxi Liang, Kongming Ma, Zhanyu |
| author_facet | Yan, Zhonghao Diao, Muxi Yang, Yuxuan Jing, Ruoyan Xu, Jiayuan Zhang, Kaizhou Yang, Lele Liu, Yanxi Liang, Kongming Ma, Zhanyu |
| contents | Accurately grounding regions of interest (ROIs) is critical for diagnosis and treatment planning in medical imaging. While multimodal large language models (MLLMs) combine visual perception with natural language, current medical-grounding pipelines still rely on supervised fine-tuning with explicit spatial hints, making them ill-equipped to handle the implicit queries common in clinical practice. This work makes three core contributions. We first define Unified Medical Reasoning Grounding (UMRG), a novel vision-language task that demands clinical reasoning and pixel-level grounding. Second, we release U-MRG-14K, a dataset of 14K samples featuring pixel-level masks alongside implicit clinical queries and reasoning traces, spanning 10 modalities, 15 super-categories, and 108 specific categories. Finally, we introduce MedReasoner, a modular framework that distinctly separates reasoning from segmentation: an MLLM reasoner is optimized with reinforcement learning, while a frozen segmentation expert converts spatial prompts into masks, with alignment achieved through format and accuracy rewards. MedReasoner achieves state-of-the-art performance on U-MRG-14K and demonstrates strong generalization to unseen clinical queries, underscoring the significant promise of reinforcement learning for interpretable medical grounding. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_08177 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MedReasoner: Reinforcement Learning Drives Reasoning Grounding from Clinical Thought to Pixel-Level Precision Yan, Zhonghao Diao, Muxi Yang, Yuxuan Jing, Ruoyan Xu, Jiayuan Zhang, Kaizhou Yang, Lele Liu, Yanxi Liang, Kongming Ma, Zhanyu Computer Vision and Pattern Recognition Artificial Intelligence Accurately grounding regions of interest (ROIs) is critical for diagnosis and treatment planning in medical imaging. While multimodal large language models (MLLMs) combine visual perception with natural language, current medical-grounding pipelines still rely on supervised fine-tuning with explicit spatial hints, making them ill-equipped to handle the implicit queries common in clinical practice. This work makes three core contributions. We first define Unified Medical Reasoning Grounding (UMRG), a novel vision-language task that demands clinical reasoning and pixel-level grounding. Second, we release U-MRG-14K, a dataset of 14K samples featuring pixel-level masks alongside implicit clinical queries and reasoning traces, spanning 10 modalities, 15 super-categories, and 108 specific categories. Finally, we introduce MedReasoner, a modular framework that distinctly separates reasoning from segmentation: an MLLM reasoner is optimized with reinforcement learning, while a frozen segmentation expert converts spatial prompts into masks, with alignment achieved through format and accuracy rewards. MedReasoner achieves state-of-the-art performance on U-MRG-14K and demonstrates strong generalization to unseen clinical queries, underscoring the significant promise of reinforcement learning for interpretable medical grounding. |
| title | MedReasoner: Reinforcement Learning Drives Reasoning Grounding from Clinical Thought to Pixel-Level Precision |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2508.08177 |