Skip to content

Latest commit

 

History

History
53 lines (35 loc) · 1.83 KB

README.md

File metadata and controls

53 lines (35 loc) · 1.83 KB

Face Recognition

Pretrained Models

WebFace4M

MS1MV2

  • ...

Benchmarks

Evaluation on 5 High Quality Image Validation Sets (LFW, CFPFP, CPLFW, CALFW, AGEDB)

Model Dataset Method LFW CFPFP CPLFW CALFW AGEDB AVG
R18 WebFace4M AdaFace 99.53 97.26 92.28 95.52 96.47 96.21
R50 MS1MV2 AdaFace 99.82 97.86 92.83 96.07 97.85 96.88
R50 WebFace4M AdaFace 99.78 98.97 94.17 95.98 97.78 97.34

Evaluation on Mixed Quality Scenario (IJBB, IJBC)

Model Dataset Method IJBB TAR@FAR=0.01% IJBC TAR@FAR=0.01%
R18 WebFace4M AdaFace 93.03 94.99
R50 MS1MV2 AdaFace 94.82 96.27
R50 WebFace4M AdaFace 95.44 96.98

Evaluation on Low Quality Scenario (IJBS)

Sur-to-Single Sur-to-Book
Model Pretrained Dataset Method Rank1 Rank5 1% Rank1 Rank5 1%
R100 MS1MV2 AdaFace 65.26 70.53 51.66 66.27 71.61 50.87
R100 WebFace4M AdaFace 70.42 75.29 58.27 70.93 76.11 58.02
Sur-to-Sur TinyFace
Model Pretrained Dataset Method Rank1 Rank5 1% Rank1 Rank5
R100 MS1MV2 AdaFace 23.74 37.47 2.50 68.21 71.54
R100 WebFace4M AdaFace 35.05 48.22 4.96 72.02 74.52
  • Sur-to-Single: Protocol comparing surveillance video (probe) to single enrollment image (gallery)
  • Sur-to-Book: Protocol comparing surveillance video (probe) to all enrollment images (gallery)
  • Sur-to-Sur: Protocol comparing surveillance video (probe) to surveillance video (gallery)

Datasets Preparation

Coming soon.