Asymmetric Rejection Loss for Fairer Face Recognition

Abstract

Face recognition performance has seen tremendous gain in the recent years, mostly due to the availability of large-scale face images dataset that can be exploited by deep neural networks to learn powerful face representations. However, recent research has shown differences of face recognition performance across different ethnic groups mostly due to the racial imbalance in the training datasets where Caucasian identities largely dominate other ethnicities. This is actually symptomatic of the under representation of non-Caucasian ethnic groups in the celebdom from which face datasets are usually gathered, rendering the acquisition of labeled data of the under-represented groups challenging. In this paper, we propose an Asymmetric Rejection Loss, which aims at making full use of unlabeled images of those under-represented groups, to reduce the racial bias of face recognition models. We view each unlabeled image as a unique class, however as we cannot guarantee that two unlabeled samples are from distinct class we exploit labeled and unlabeled data in an asymmetric manner in our loss formalism. Extensive experiments show our method's strength in mitigating racial bias, outperforming state-ofthe-art semi-supervision methods. Performance on the under-represented ethnicity groups increases while that on well-represented group is nearly unchanged.

Mitigating Face Recognition Bias via Group Adaptive Classifier

“…Accordingly, a disentangled representation is learned to de-bias both FR and demographic attribute estimation [19]. Other studies address the bias issue in FR by leveraging unlabeled faces to improve the performance in groups with fewer samples [55,67]. Wang et al [66] propose skewness-aware reinforcement learning to mitigate racial bias in FR.…”