Medical Image Classification on MammoBench (20% hold-out split)
62.7Balanced AccuracyDSBR
Evaluation results
| Method | Links | |
|---|---|---|
| DSBRBackbone=ResNet2026.06 | 62.7 | |
| DeYOBackbone=ResNet2026.06 | 60.9 | |
| ResNetBackbone=ResNet2026.06 | 59.4 | |
| SARBackbone=ResNet2026.06 | 59.4 | |
| TentBackbone=ResNet2026.06 | 58.8 | |
| ROIDBackbone=ResNet2026.06 | 56.6 | |
| DeYOBackbone=ViT2026.06 | 55.2 | |
| DSBRBackbone=ViT2026.06 | 55.2 | |
| ROIDBackbone=ViT2026.06 | 54.5 | |
| ViTBackbone=ViT2026.06 | 54.4 | |
| TentBackbone=ViT2026.06 | 54.4 | |
| SARBackbone=ViT2026.06 | 54.4 |