Diagnosis classification on Derm7pt (test)
77.3F1ViP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ViPBackbone=Mean across 3 backbones (ViT-B/16, ViT-L/14, ResNet-50), Evaluation Protocol=zero-shot2024.08 | 77.3 | — | 81.11 | |
| CoOpBackbone=Mean across 3 backbones (ViT-B/16, ViT-L/14, ResNet-50)2024.08 | 73.28 | — | 78.23 | |
| MaPLeBackbone=Mean across 3 backbones (ViT-B/16, ViT-L/14, ResNet-50)2024.08 | 72.5 | — | 79.03 | |
| DModality=Dermoscopic2022.03 | 71.6 | 63.5 | — | |
| KgCoOpBackbone=Mean across 3 backbones (ViT-B/16, ViT-L/14, ResNet-50)2024.08 | 70.93 | — | 77.26 | |
| D + CModality=Dermoscopic + Clinical2022.03 | 68.8 | 48.4 | — | |
| CoCoOpBackbone=Mean across 3 backbones (ViT-B/16, ViT-L/14, ResNet-50)2024.08 | 66.47 | — | 76.68 | |
| SupervisedBackbone=Mean across 3 backbones (ViT-B/16, ViT-L/14, ResNet-50)2024.08 | 62.36 | — | 72.77 | |
| G(C) + CModality=Clinical, Guidance=Learned mapping + Original clinical inputs2022.03 | 61.8 | 44.7 | — | |
| G(C)Modality=Clinical, Guidance=Learned mapping from clinical features to superior dermoscopic features2022.03 | 61.6 | 33 | — | |
| CModality=Clinical2022.03 | 60.4 | 43.8 | — | |
| BayesianBackbone=Mean across 3 backbones (ViT-B/16, ViT-L/14, ResNet-50)2024.08 | 55.61 | — | 69.21 |