Research map / Vision-language and multimodal / Multimodal LLMs
Skin-R1: Clinical Knowledge-Guided Dermatological Diagnosis Using Vision-Language Models
Zehao Liu, Weijieying Ren, Jipeng ZHANG, Tianxiang Zhao, Jingxi Zhu, Xiaoting Li, Vasant Honavar
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add textbook-trajectory SFT before RL (ablation) | Skin-R1 (Qwen2.5-VL-7B) · Disease diagnosis, ID avg | accuracy: 0.5843 -> 0.6385 | Improved |
| add textbook-trajectory SFT before RL (ablation) | Skin-R1 (Qwen2.5-VL-7B) · Disease diagnosis, OOD avg | accuracy: 0.5674 -> 0.7171 | Improved |
| add textbook-trajectory SFT before RL (ablation) | Skin-R1 (Qwen2.5-VL-7B) · Lesion condition classification, ID avg | accuracy: 0.4602 -> 0.6928 | Improved |
| SFT on SkinRationale trajectories (ablation) | Qwen2.5-VL-7B · Disease diagnosis, ID avg | accuracy: 0.4367 -> 0.4743 | Improved |
| SFT on SkinRationale trajectories (ablation) | Qwen2.5-VL-7B · Disease diagnosis, OOD avg | accuracy: 0.5027 -> 0.5153 | Mixed |
| SFT on SkinRationale trajectories (ablation) | Qwen2.5-VL-7B · Lesion condition classification, ID avg | accuracy: 0.3698 -> 0.6271 | Improved |
| SFT on SkinRationale trajectories (ablation) | Qwen2.5-VL-7B · Valid answer rate, ID diagnosis | invalid-answer rate: about 30% -> 2% invalid | Improved |
| add granularity-wise hierarchical reward (vs standard binary reward) (ablation) | Skin-R1 · Disease diagnosis, OOD avg | accuracy: 0.6386 -> 0.7171 | Improved |
| add granularity-wise hierarchical reward (vs standard binary reward) (ablation) | Skin-R1 · Disease diagnosis, ID avg | accuracy: 0.6379 -> 0.6385 | No clear effect |
| add granularity-wise hierarchical reward (vs standard binary reward) (ablation) | Skin-R1 · Lesion condition classification, ID avg | accuracy: 0.6658 -> 0.6928 | Improved |
| train RL longer (500 -> 1,500 steps) (ablation) | Skin-R1 · Disease diagnosis, ID avg | accuracy: 0.5875 -> 0.6385 | Improved |
| train RL longer (500 -> 1,500 steps) (ablation) | Skin-R1 · Disease diagnosis, OOD avg | accuracy: 0.6742 -> 0.7171 | Improved |
Closest papers by meaning
- Doctor Approved: Generating Medically Accurate Skin Disease Images through AI-Expert Feedback (NeurIPS 2025 · similarity 0.90)
- DermaCon-IN: A Multiconcept-Annotated Dermatological Image Dataset of Indian Skin Disorders for Clinical AI Research (NeurIPS 2025 · similarity 0.87)
- Patho-R1: A Multimodal Reinforcement Learning-Based Pathology Expert Reasoner (AAAI 2026 · similarity 0.87)
- MICA: Towards Explainable Skin Lesion Diagnosis via Multi-Level Image-Concept Alignment (AAAI 2024 · similarity 0.86)
- MedReasoner: Reinforcement Learning Drives Reasoning Grounding from Clinical Thought to Pixel-Level Precision (AAAI 2026 · similarity 0.85)
- MedTVT-R1: A Multimodal LLM Empowering Medical Reasoning and Diagnosis (CVPR 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
