Medical Report Generation on MIMIC-CXR findings
34.02BLEU-1DeepMedix-R1
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DeepMedix-R12025.09 | 34.02 | 21.49 | 14.79 | 10.59 | 28.95 | 32.92 | 23.87 | 31.26 | 51.89 | 48.21 | 57.26 | 32.3 | |
| LLaVA-Rad2025.09 | 29.11 | 17.57 | 11.89 | 8.46 | 25 | 29.28 | 20.77 | 31.54 | 47.53 | 45.92 | 55.35 | 29.31 | |
| DeepMedix2025.09 | 24.45 | 15.89 | 11.47 | 8.71 | 22.91 | 28.7 | 18.32 | 22.71 | 39.56 | 34.76 | 44.77 | 24.75 | |
| HuatuoGPT-Vision2025.09 | 22.44 | 11.33 | 6.31 | 3.9 | 23.61 | 19.98 | 11.27 | 22.22 | 35.32 | 35.43 | 40.99 | 21.16 | |
| MedGemma2025.09 | 17.52 | 9.6 | 5.86 | 3.77 | 26.13 | 21.22 | 18.64 | 33.47 | 50.41 | 46.16 | 51.83 | 25.87 | |
| DeepSeek-VL-7B2025.09 | 13.56 | 5.66 | 3 | 1.94 | 16.73 | 17.3 | 6.19 | 13 | 24.44 | 11.18 | 17.1 | 11.83 | |
| Qwen2.5-VL-32B2025.09 | 8.87 | 4.4 | 2.35 | 1.36 | 21.27 | 15.38 | 9.21 | 22.6 | 34.61 | 32.58 | 34.76 | 17.04 | |
| Qwen2.5-VL-7B2025.09 | 7.32 | 3.67 | 2 | 1.17 | 19.32 | 14.34 | 9.3 | 13.27 | 21.37 | 16.43 | 17.69 | 11.44 | |
| Qwen2.5-VL-7Bthinking=true2025.09 | 6.15 | 3.04 | 1.97 | 1.47 | 7.64 | 11.06 | 4.93 | 11.98 | 17.41 | 14.92 | 15.09 | 8.7 | |
| CheXagent2025.09 | 2.77 | 1.7 | 1.34 | 1.15 | 4.24 | 6.58 | 9.65 | 12.75 | 18.18 | 19.35 | 24.79 | 9.32 |