Image Classification on DamageNet
70.42Top-1 AccuracyMAE-H + DAT
Evaluation Results
| Method | Links | |
|---|---|---|
| MAE-H + DAT2022.09 | 70.42 | |
| MAE-H2022.09 | 64.36 | |
| AugReg-ViT + DAT2022.09 | 45.7 | |
| AugReg-ViT2022.09 | 45.24 | |
| DrViT2022.09 | 44.91 | |
| ViT2022.09 | 28.99 | |
| DeepAugment + Augmix + DAT2022.09 | 22.86 | |
| DeepAugment + Augmix2022.09 | 19.6 | |
| ResNet50 + DAT2022.09 | 14.42 | |
| DeepAugment2022.09 | 11.94 | |
| EMD-CorrFeatures=ImageNet, Training set=ImageNet2022.07 | 8.16 | |
| CHM-CorrFeatures=ImageNet, Training set=ImageNet2022.07 | 8.1 | |
| kNNFeatures=ImageNet, Training set=ImageNet2022.07 | 7.59 | |
| ResNet502022.09 | 5.94 | |
| ResNet-50Features=ImageNet, Training set=ImageNet2022.07 | 5.93 |