Federated Unsupervised Domain Adaptation for Face Recognition

TL;DR

This paper introduces FedFR, a federated unsupervised domain adaptation method for face recognition that preserves privacy while improving target domain performance through clustering and domain constraints.

Contribution

FedFR combines clustering-based domain adaptation with federated learning, introducing a new domain constraint loss to enhance face recognition across domains without sharing sensitive data.

Findings

FedFR outperforms baseline methods by 3% to 14% on various metrics.

Enhanced clustering improves pseudo label quality for unlabeled target data.

Domain constraint loss regularizes source training in federated setting.

Abstract

Given labeled data in a source domain, unsupervised domain adaptation has been widely adopted to generalize models for unlabeled data in a target domain, whose data distributions are different. However, existing works are inapplicable to face recognition under privacy constraints because they require sharing of sensitive face images between domains. To address this problem, we propose federated unsupervised domain adaptation for face recognition, FedFR. FedFR jointly optimizes clustering-based domain adaptation and federated learning to elevate performance on the target domain. Specifically, for unlabeled data in the target domain, we enhance a clustering algorithm with distance constrain to improve the quality of predicted pseudo labels. Besides, we propose a new domain constraint loss (DCL) to regularize source domain training in federated learning. Extensive experiments on a newly constructed benchmark demonstrate that FedFR outperforms the baseline and classic methods on the target domain by 3% to 14% on different evaluation metrics.