Face Recognition on IJB-S V2S
72.54Rank-1 AccuracyAdaFace+Ours
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AdaFace+OursLabeled Train Data=WebFace12M, Backbone=IResNet-1002022.07 | 72.54 | 77.59 | 60.94 | 66.02 | |
| AdaFaceLabeled Train Data=WebFace12M, Backbone=IResNet-1002022.07 | 71.35 | 76.24 | 59.4 | 66.34 | |
| ArcFace+OursLabeled Train Data=MS1MV2-*, Backbone=ResNet-502022.07 | 63.86 | 69.95 | 47.86 | 56.44 | |
| ArcFace+Ours*Labeled Train Data=MS1MV2-*, Backbone=ResNet-502022.07 | 61.69 | 68.33 | 43.99 | 53.34 | |
| Shi et al.Labeled Train Data=MS1MV2-*, Backbone=ResNet-502022.07 | 59.29 | 66.91 | 39.92 | 50.49 | |
| ArcFaceLabeled Train Data=MS1MV2-*, Backbone=ResNet-502022.07 | 58.78 | 66.4 | 40.99 | 50.45 | |
| MARNLabeled Train Data=MS1M-*, Backbone=ResNet-642022.07 | 58.14 | 64.11 | 21.47 | — | |
| C-FANLabeled Train Data=MS1M-*, Backbone=ResNet-642022.07 | 50.82 | 61.16 | 16.44 | 24.19 | |
| ArcFaceLabeled Train Data=MS1MV2, Backbone=ResNet-502022.07 | 50.39 | 60.42 | 32.39 | 42.99 | |
| PFELabeled Train Data=MS1M-*, Backbone=ResNet-642022.07 | 50.16 | 58.33 | 31.88 | 35.33 |