To reproduce the figures and tables in the notebook, please download everything (model, code, data and meta info) from here: Dropbox or Baidu Cloud
Updated Meta data (1:1 and 1:N): IJB-B Dropbox and IJB-C Dropbox
Please apply for the IJB-B and IJB-C by yourself and strictly follow their distribution licenses.
Aknowledgement
Great thanks for Weidi Xie's instruction [2,3,4,5] to evaluate ArcFace [1] on IJB-B[6] and IJB-C[7] (1:1 protocol).
Great thanks for Yuge Huang's code [8] to evaluate ArcFace [1] on IJB-B[6] and IJB-C[7] (1:N protocol).
Reference
[1] Jiankang Deng, Jia Guo, Niannan Xue, Stefanos Zafeiriou. Arcface: Additive angular margin loss for deep face recognition[J]. arXiv:1801.07698, 2018.
[2] https://github.com/ox-vgg/vgg_face2.
[3] Qiong Cao, Li Shen, Weidi Xie, Omkar M Parkhi, Andrew Zisserman. VGGFace2: A dataset for recognising faces across pose and age. FG, 2018.
[4] Weidi Xie, Andrew Zisserman. Multicolumn Networks for Face Recognition. BMVC 2018.
[5] Weidi Xie, Li Shen, Andrew Zisserman. Comparator Networks. ECCV, 2018.
[6] Whitelam, Cameron, Emma Taborsky, Austin Blanton, Brianna Maze, Jocelyn C. Adams, Tim Miller, Nathan D. Kalka et al. IARPA Janus Benchmark-B Face Dataset. CVPR Workshops, 2017.
[7] Maze, Brianna, Jocelyn Adams, James A. Duncan, Nathan Kalka, Tim Miller, Charles Otto, Anil K. Jain et al. IARPA Janus Benchmark–C: Face Dataset and Protocol. ICB, 2018.
[8] Yuge Huang, Pengcheng Shen, Ying Tai, Shaoxin Li, Xiaoming Liu, Jilin Li, Feiyue Huang, Rongrong Ji. Distribution Distillation Loss: Generic Approach for Improving Face Recognition from Hard Samples. arXiv:2002.03662.