Multi-label classification on Clinical NCCT dataset
74Micro SENBrain-Adapter
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Brain-AdapterLoss Function=Asymmetric Loss (ASL)2026.06 | 74 | 86.5 | 88.7 | 77.8 | 7.9 | |
| Swin-BModel Category=3D Volumetric Models2026.06 | 61.2 | 80.2 | 77.8 | 53.5 | 13.1 | |
| ViT-BModel Category=3D Volumetric Models2026.06 | 56.7 | 78 | 75.9 | 53.4 | 24 | |
| ABMILModel Category=2D MIL Models, Pre-training Strategy=BiomedCLIP Pre-trained2026.06 | 40 | 96.5 | 86 | 75 | 8.8 | |
| TransMILModel Category=2D MIL Models, Pre-training Strategy=Trained from Scratch2026.06 | 32.8 | 95.6 | 81.9 | 74.9 | 10.3 | |
| Brain-AdapterLoss Function=standard BCE loss2026.06 | 28.5 | 98.2 | 87.6 | 77.3 | 8.4 | |
| ABMILModel Category=2D MIL Models, Pre-training Strategy=Trained from Scratch2026.06 | 28.4 | 98.1 | 79.7 | 56.3 | 9.4 | |
| Mean PoolingModel Category=2D MIL Models, Pre-training Strategy=Trained from Scratch2026.06 | 26.6 | 97.7 | 77.5 | 49.3 | 9.4 | |
| TransMILModel Category=2D MIL Models, Pre-training Strategy=BiomedCLIP Pre-trained2026.06 | 15.5 | 99.3 | 85.7 | 69.8 | 8.6 | |
| Mean PoolingModel Category=2D MIL Models, Pre-training Strategy=BiomedCLIP Pre-trained2026.06 | 10.4 | 99.6 | 84 | 66.1 | 8.7 |