Medical Image Classification on HAM10000 (test)
0.952AUCCUPID
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CUPIDBackbone=ResNet182026.03 | 0.952 | 82.1 | |
| IGRUEBackbone=ResNet182026.03 | 0.952 | 82.1 | |
| Rate-inBackbone=ResNet182026.03 | 0.952 | 82.1 | |
| MC DropoutBackbone=ResNet18, Drop rate=0.03, Stochastic forward passes=102026.03 | 0.946 | 76.5 | |
| BNNBackbone=ResNet182026.03 | 0.929 | 74.2 | |
| DECBackbone=ResNet182026.03 | 0.922 | 76.8 | |
| PostNetBackbone=ResNet182026.03 | 0.865 | 66.4 |