Medical Image Segmentation on Site A
84.1Dice ScoreMCDRL
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MCDRLBackbone=ViT-L/14, Pre-trained weights=CLIP2025.08 | 84.1 | 76.8 | 95.9 | |
| MCDRLBackbone=ViT-B/16, Pre-trained weights=CLIP2025.08 | 82.6 | 75.4 | 95.1 | |
| BiomedCoOpBackbone=ViT-L/14, Pre-trained weights=CLIP2025.08 | 81.7 | 74.6 | 94.6 | |
| MCDRLBackbone=ResNet-50, Pre-trained weights=CLIP2025.08 | 81.5 | 74.2 | 94.3 | |
| StyLIPBackbone=ViT-L/14, Pre-trained weights=CLIP2025.08 | 80.6 | 73.5 | 94.1 | |
| BiomedCoOpBackbone=ViT-B/16, Pre-trained weights=CLIP2025.08 | 80.3 | 73.4 | 93.8 | |
| BiomedCoOpBackbone=ResNet-50, Pre-trained weights=CLIP2025.08 | 79.2 | 72.6 | 93.1 | |
| StyLIPBackbone=ViT-B/16, Pre-trained weights=CLIP2025.08 | 79.1 | 72.2 | 93.2 | |
| StyLIPBackbone=ResNet-50, Pre-trained weights=CLIP2025.08 | 78.1 | 71.1 | 92.5 | |
| BaselineBackbone=ViT-L/14, Pre-trained weights=CLIP2025.08 | 78.1 | 70.7 | 93 | |
| BaselineBackbone=ViT-B/16, Pre-trained weights=CLIP2025.08 | 76.4 | 69.3 | 92.1 | |
| BaselineBackbone=ResNet-50, Pre-trained weights=CLIP2025.08 | 75.3 | 68.2 | 91.2 |