Image Classification on Facescrub (train)
0.0012Dec.-based MSENo_defense
Evaluation results
| Method | Links | ||
|---|---|---|---|
| No_defenseBackbone=VGG112025.03 | 0.0012 | 0.0012 | |
| BottleneckBackbone=VGG112025.03 | 0.0025 | 0.0026 | |
| Bottleneck+CEMBackbone=VGG112025.03 | 0.0036 | 0.0038 | |
| DistCorrBackbone=VGG112025.03 | 0.0038 | 0.0041 | |
| DistCorr+CEMBackbone=VGG112025.03 | 0.0048 | 0.0069 | |
| DropoutBackbone=VGG112025.03 | 0.0052 | 0.0054 | |
| Noise_NopeekBackbone=VGG112025.03 | 0.0052 | 0.0053 | |
| Dropout+CEMBackbone=VGG112025.03 | 0.0074 | 0.0076 | |
| Noise_Nopeek+CEMBackbone=VGG112025.03 | 0.0076 | 0.0078 | |
| ResSFLBackbone=VGG112025.03 | 0.0094 | 0.0111 | |
| PATROLBackbone=VGG112025.03 | 0.0099 | 0.0114 | |
| Noise_ARLBackbone=VGG112025.03 | 0.0122 | 0.0132 | |
| ResSFL+CEMBackbone=VGG112025.03 | 0.0128 | 0.0143 | |
| PATROL+CEMBackbone=VGG112025.03 | 0.0166 | 0.0184 | |
| Noise_ARL+CEMBackbone=VGG112025.03 | 0.0182 | 0.0212 |