List your company
ResearchFormerly WizWand

Image Classification on DamageNet

70.42Top-1 AccuracyMAE-H + DAT
3.350420.762738.17555.5873Jul 26, 2022Aug 3, 2022Aug 12, 2022Aug 21, 2022Aug 29, 2022Sep 7, 2022Sep 16, 2022
Updated 8mo ago

Evaluation results

MethodLinks
MAE-H + DAT2022.0970.42
MAE-H2022.0964.36
AugReg-ViT + DAT2022.0945.7
AugReg-ViT2022.0945.24
DrViT2022.0944.91
ViT2022.0928.99
DeepAugment + Augmix + DAT2022.0922.86
DeepAugment + Augmix2022.0919.6
ResNet50 + DAT2022.0914.42
DeepAugment2022.0911.94
EMD-CorrFeatures=ImageNet, Training set=ImageNet2022.078.16
CHM-CorrFeatures=ImageNet, Training set=ImageNet2022.078.1
kNNFeatures=ImageNet, Training set=ImageNet2022.077.59
ResNet502022.095.94
ResNet-50Features=ImageNet, Training set=ImageNet2022.075.93