Medical Imaging Report Generation on Med-HallMark
64.49BertScoremPLUG-Owl2
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| mPLUG-Owl22024.06 | 64.49 | 40.11 | 32 | 13.84 | 28.5 | 6.79 | 0.81 | |
| XrayGPT2024.06 | 62.62 | 25.96 | 27.94 | 6.59 | 22.15 | 3.26 | 0.79 | |
| LLaVA1.5Model size=13b2024.06 | 47.96 | 11.8 | 18.35 | 2.4 | 14.49 | 0.67 | 0.81 | |
| LLaVA1.5Model size=7b2024.06 | 47.93 | 11.24 | 18.77 | 2.78 | 14.64 | 0.72 | 0.78 | |
| InstructBLIP2024.06 | 47.49 | 13.98 | 17.56 | 2.31 | 13.6 | 0.73 | 0.84 | |
| InstructBLIPModel size=13b2024.06 | 47.47 | 13.93 | 17.54 | 2.34 | 13.61 | 0.74 | 0.84 | |
| mini-gpt42024.06 | 46.43 | 10.27 | 15.37 | 1.75 | 12.63 | 0.53 | 0.88 | |
| RadFM2024.06 | 41.18 | 3.41 | 5.88 | 0.76 | 4.79 | 0.05 | — | |
| LLaVA-MedFine-tuned dataset=Pathvqa2024.06 | 37.67 | 1.25 | 2.24 | 0.05 | 2.06 | 0.09 | — | |
| LLaVA-MedFine-tuned dataset=Rad2024.06 | 33.28 | 2.33 | 2.73 | 0.32 | 2.45 | 0.02 | — | |
| BLIP22024.06 | 33.05 | 4.88 | 10.17 | 1.49 | 7.66 | 0.15 | — | |
| LLaVA-MedFine-tuned dataset=Slake2024.06 | 30.49 | 0.61 | 0.27 | 0 | 0.27 | 0.14 | — |