Face Verification on LFW (unrestricted-labeled-outside-data protocol 14)
99.85AccuracyAFRN
Evaluation results
| Method | Links | |
|---|---|---|
| AFRNTrain Data=MS1MV2, Backbone=ResNet100, Venue=ICCV192022.04 | 99.85 | |
| BroadFaceTrain Data=MS1MV2, Backbone=ResNet100, Venue=ECCV202022.04 | 99.85 | |
| ArcFaceTrain Data=MS1MV2, Backbone=ResNet100, Margin (m)=0.50, Venue=CVPR192022.04 | 99.83 | |
| MagFaceTrain Data=MS1MV2, Backbone=ResNet100, Venue=CVPR212022.04 | 99.83 | |
| VPL-ArcFaceTrain Data=MS1MV3, Backbone=ResNet100, Venue=CVPR212022.04 | 99.83 | |
| AdaFaceTrain Data=MS1MV3, Backbone=ResNet100, Margin (m)=0.4, Venue=CVPR222022.04 | 99.83 | |
| ArcFace*Train Data=WebFace4M, Backbone=ResNet100, Venue=CVPR192022.04 | 99.83 | |
| SCF-ArcFaceTrain Data=MS1MV2, Backbone=ResNet100, Venue=CVPR212022.04 | 99.82 | |
| AdaFaceTrain Data=MS1MV2, Backbone=ResNet100, Margin (m)=0.4, Venue=CVPR222022.04 | 99.82 | |
| CosFaceTrain Data=MS1MV2, Backbone=ResNet100, Margin (m)=0.35, Venue=CVPR182022.04 | 99.81 | |
| MV-SoftmaxTrain Data=MS1MV2, Backbone=ResNet100, Venue=AAAI202022.04 | 99.8 | |
| CurricularFaceTrain Data=MS1MV2, Backbone=ResNet100, Venue=CVPR202022.04 | 99.8 | |
| AdaFaceTrain Data=WebFace4M, Backbone=ResNet100, Margin (m)=0.4, Venue=CVPR222022.04 | 99.8 | |
| L2-S (RX101)Images=3.7M, #nets=1, One loss=Yes, Backbone=ResNeXt-101, Loss=L2-softmax2017.03 | 99.78 | |
| URLTrain Data=MS1MV2, Backbone=ResNet100, Venue=CVPR202022.04 | 99.78 | |
| Baidu# of CNNs=10, Dataset Info=1.2M, 1.8K2017.03 | 99.77 | |
| L2-S (R101)Images=3.7M, #nets=1, One loss=Yes, Backbone=ResNet-101, Loss=L2-softmax2017.03 | 99.67 | |
| FaceNet# of CNNs=1, Dataset Info=200M, 8M2017.03 | 99.63 | |
| FaceNetImages=200M, #nets=1, One loss=Yes2017.03 | 99.63 | |
| DeepVisage# of CNNs=1, Dataset Info=4.48M, 62K2017.03 | 99.62 | |
| DeepVisageImages=4.48M, #nets=1, One loss=Yes2017.03 | 99.62 | |
| L2-S (FR)Images=3.7M, #nets=1, One loss=Yes, Backbone=Face-Resnet, Loss=L2-softmax2017.03 | 99.6 | |
| Sparse ConvNet# of CNNs=25, Dataset Info=0.29M, 12K2017.03 | 99.55 | |
| DeepID3# of CNNs=25, Dataset Info=0.29M, 12K2017.03 | 99.53 | |
| Megvii# of CNNs=4, Dataset Info=5M, 0.2M2017.03 | 99.5 | |
| LF-CNNs# of CNNs=25, Dataset Info=0.7M, 17.2K2017.03 | 99.5 | |
| DeepID2+# of CNNs=25, Dataset Info=0.29M, 12K2017.03 | 99.47 | |
| DeepID-2+Images=-, #nets=25, One loss=No2017.03 | 99.47 | |
| SphereFaceImages=0.5M, #nets=1, One loss=Yes2017.03 | 99.42 | |
| Center Loss# of CNNs=1, Dataset Info=0.7M, 17.2K2017.03 | 99.28 | |
| Wen et al.Images=0.7M, #nets=1, One loss=No2017.03 | 99.28 | |
| Baidu# of CNNs=1, Dataset Info=1.2M, 1.8K2017.03 | 99.13 | |
| BaiduImages=1.3M, #nets=1, One loss=No2017.03 | 99.13 | |
| MM-DFR# of CNNs=8, Dataset Info=0.49M, 10.57K2017.03 | 99.02 | |
| softmax(FR)Images=3.7M, #nets=1, One loss=Yes, Backbone=Face-Resnet2017.03 | 99 | |
| VGG Face# of CNNs=1, Dataset Info=2.6M, 2.6K2017.03 | 98.95 | |
| VGG FaceImages=2.6M, #nets=1, One loss=No2017.03 | 98.95 | |
| MFM-CNN# of CNNs=1, Dataset Info=5.1M, 79K2017.03 | 98.8 | |
| VIPLFaceNet# of CNNs=1, Dataset Info=0.49M, 10.57K2017.03 | 98.6 | |
| Webscale# of CNNs=4, Dataset Info=4.5M, 55K2017.03 | 98.37 | |
| AAL# of CNNs=1, Dataset Info=0.49M, 10.57K2017.03 | 98.3 | |
| FSS# of CNNs=9, Dataset Info=0.49M, 10.57K2017.03 | 98.2 | |
| Face-Aug-Pose-Syn# of CNNs=1, Dataset Info=2.4M, 10.57K2017.03 | 98.06 | |
| CASIA-Webface# of CNNs=1, Dataset Info=0.49M, 10.57K2017.03 | 97.73 | |
| Unconstrained FV# of CNNs=1, Dataset Info=0.49M, 10.5K2017.03 | 97.45 | |
| Deepface# of CNNs=3, Dataset Info=4.4M, 4K2017.03 | 97.35 | |
| Deep FaceImages=4M, #nets=3, One loss=No2017.03 | 97.35 |