Medical Image Segmentation on GAMMA relabeled
88.6Vessel IoURetSAM
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| RetSAMModel Architecture=RetSAM, Evaluation Protocol=Finetune2026.01 | 88.6 | 98.4 | 95.5 | 73.2 | 0.682 | |
| RetSAMModel Architecture=RetSAM, Evaluation Protocol=Linear2026.01 | 72.5 | 95.4 | 61.7 | 48.2 | 0.45 | |
| SAM3Model Architecture=SAM32026.01 | 72 | 96.6 | 74.8 | 22.5 | 0.402 | |
| SAM2Model Architecture=SAM22026.01 | 71 | 96.1 | 73.7 | 22 | 0.394 | |
| SAM2-UNetModel Architecture=SAM2-UNet2026.01 | 60 | 96 | 89.3 | — | — | |
| RetFoundModel Architecture=RetFound2026.01 | 58 | 94.9 | 78.4 | — | — | |
| MedSAM2Model Architecture=MedSAM22026.01 | 43.9 | 94.3 | 77.8 | — | — | |
| MedSAMModel Architecture=MedSAM2026.01 | 9.5 | 92.4 | 76.3 | 0.1 | — | |
| Med-SAM-AdapterModel Architecture=Med-SAM-Adapter2026.01 | 3.7 | 90.5 | 69.8 | — | — |