Abnormal classification on VinDr-Mammo (test)
0.841AUCDisAsymNet
Evaluation results
| Method | Links | |
|---|---|---|
| DisAsymNetBackbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.841 | |
| AsyC & AsyDBackbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.833 | |
| AsyDBackbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.827 | |
| AsyD & SynBackbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.825 | |
| AsyCBackbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.809 | |
| HAMBackbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.803 | |
| Late fusionBackbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.803 | |
| CVTBackbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.797 | |
| Wang et al.Backbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.797 | |
| ResNet-18Backbone=ResNet-18, Pre-training=ImageNet, Image Size=1024x5122023.07 | 0.782 |