Image Classification on ISIC-DICM-17K 1:1 class-balanced (test)
0.8822AUCSemCovNet
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SemCovNet2026.02 | 0.8822 | 0.8098 | 0.4138 | 0.7893 | 0.7855 | 0.0421 | |
| ENet-B0Backbone=ENet-B02026.02 | 0.8612 | 0.773 | 0.3802 | 0.76 | 0.7646 | 0.0613 | |
| ENet-B0+CBLBackbone=ENet-B0, Loss=Class-Balanced Loss2026.02 | 0.8612 | 0.773 | 0.3802 | 0.76 | 0.7646 | 0.0613 | |
| MONET2026.02 | 0.8598 | 0.79 | 0.403 | 0.7653 | 0.765 | 0.0661 | |
| GroupDROStrategy=Distributionally Robust Optimization2026.02 | 0.8557 | 0.7632 | 0.3881 | 0.7479 | 0.7542 | 0.1118 | |
| CLIPBackbone=CLIP2026.02 | 0.8525 | 0.7839 | 0.3855 | 0.7546 | 0.756 | 0.0565 | |
| ViTBackbone=ViT2026.02 | 0.8098 | 0.7182 | 0.2937 | 0.6675 | 0.6725 | 0.041 | |
| ENet-B0+ASLBackbone=ENet-B0, Loss=Asymmetric Loss2026.02 | 0.7543 | 0.6569 | 0.2605 | 0.6591 | 0.6634 | 0.2047 |