RamFace: Race Adaptive Margin Based Face Recognition for Racial Bias Mitigation

Zhanjia Yang, Xiangping Zhu, Changyuan Jiang, Wenshuang Liu, Linlin Shen

Research output: Chapter in Book/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

Recent studies show that there exist significant racial bias among state-of-The-Art (SOTA) face recognition algorithms, i.e., the accuracy for Caucasian is consistently higher than that for other races like African and Asian. To mitigate racial bias, we propose the race adaptive margin based face recognition (RamFace) model, designed under the multi-Task learning framework with the race classification as the auxiliary task. The experiments show that the race classification task can enforce the model to learn the racial features and thus improve the discriminability of the extracted feature representations. In addition, a racial bias robust loss function, i.e., race adaptive margin loss, is proposed such that different optimal margins can be automatically derived for different races in training the model, which further mitigates the racial bias. The experimental results show that on RFW dataset, our model not only achieves SOTA face recognition accuracy but also mitigates the racial bias problem. Besides, RamFace is also tested on several public face recognition evaluation benchmarks, i.e., LFW, CPLFW and CALFW, and achieves better performance than the commonly used face recognition methods, which justifies the generalization capability of RamFace.

Original languageEnglish
Title of host publication2021 IEEE International Joint Conference on Biometrics, IJCB 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665437806
DOIs
Publication statusPublished - 4 Aug 2021
Externally publishedYes
Event2021 IEEE International Joint Conference on Biometrics, IJCB 2021 - Shenzhen, China
Duration: 4 Aug 20217 Aug 2021

Publication series

Name2021 IEEE International Joint Conference on Biometrics, IJCB 2021

Conference

Conference2021 IEEE International Joint Conference on Biometrics, IJCB 2021
Country/TerritoryChina
CityShenzhen
Period4/08/217/08/21

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Biomedical Engineering

Fingerprint

Dive into the research topics of 'RamFace: Race Adaptive Margin Based Face Recognition for Racial Bias Mitigation'. Together they form a unique fingerprint.

Cite this