List your company
ResearchFormerly WizWand

Face Recognition on IJB-A (test)

94.1TAR @ FAR=0.01NAN
71.7477.54583.3589.155Mar 24, 2017
Updated 8mo ago

Evaluation results

MethodLinks
NANExternal training=true, CNN fine-tuning=false2017.0394.188.1
TAExternal training=true, CNN fine-tuning=false2017.0393.983.6
All-In-One + TPEExternal training=true, CNN fine-tuning=false2017.0392.282.3
TPEExternal training=true, CNN fine-tuning=false2017.039081.3
All-In-OneExternal training=false, CNN fine-tuning=false2017.0389.378.7
DeepVisageExternal training=false, CNN fine-tuning=false2017.0388.782.4
Face-Aug-Pose-SynExternal training=false, CNN fine-tuning=false2017.0388.672.5
TPEExternal training=false, CNN fine-tuning=false2017.0387.176.6
Unconstrained FVExternal training=true, CNN fine-tuning=true2017.0383.8—
Pose aware FRExternal training=false, CNN fine-tuning=false2017.0382.665.2
VGG FaceExternal training=false, CNN fine-tuning=false2017.0380.560.4
TSEExternal training=true, CNN fine-tuning=true2017.037959
Deep MultiposeExternal training=false, CNN fine-tuning=false2017.0378.7—
End-To-EndExternal training=false, CNN fine-tuning=true2017.0378.7—
FSSExternal training=false, CNN fine-tuning=true2017.0372.951
Sparse ConvNetExternal training=true, CNN fine-tuning=false2017.0372.646