Medical Diagnosis on MAU (test)
53DL ScoreUMed-LVLM
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| UMed-LVLMtraining_dataset=MAU2025.01 | 53 | 99 | 98 | 37 | 86 | 75 | |
| GPT-4V w/ bboxaccess=Closed-Source Model, abnormal_region_location_info=true2025.01 | 50 | 95 | 95 | 32 | 81 | 72 | |
| MedVInt (SFT&PPO)training_dataset=MAU, strategy=SFT & Proximal Policy Optimization2025.01 | 44 | 94 | 95 | 30 | 80 | 69 | |
| MedVInt (SFT)training_dataset=MAU, strategy=Supervised Fine-Tuning2025.01 | 42 | 93 | 93 | 28 | 78 | 67 | |
| MedVInt2025.01 | 29 | 11 | 27 | 8 | 9 | 17 | |
| Med-Flamingo2025.01 | 27 | 15 | 15 | 9 | 2 | 14 | |
| GPT-4Vaccess=Closed-Source Model2025.01 | 27 | 36 | 53 | 18 | 19 | 31 | |
| LLaVA-Med2025.01 | 22 | 4 | 12 | 3 | 1 | 8 | |
| LLAVA2025.01 | 20 | 0 | 4 | 0 | 0 | 5 | |
| XrayGPT2025.01 | 18 | 12 | 2 | 7 | 6 | 9 | |
| Qwen-VL2025.01 | 13 | 0 | 1 | 0 | 0 | 3 | |
| mPLUG-Owl2025.01 | 5 | 0 | 1 | 0 | 0 | 1 | |
| MiniGPT-42025.01 | 2 | 0 | 2 | 0 | 0 | 1 |