FocusFace: Multi-task Contrastive Learning for Masked Face Recognition

TL;DR

FocusFace introduces a multi-task contrastive learning architecture that significantly improves masked face recognition accuracy, outperforming existing methods especially in masked-masked verification, while reducing training costs when combined with current models.

Contribution

The paper presents a novel multi-task contrastive learning framework for masked face recognition that can be trained from scratch or integrated with existing models without losing their original capabilities.

Findings

Outperforms all compared methods in masked-masked face verification.

Achieves comparable results to state-of-the-art in unmasked scenarios.

Reduces training computational costs when used with existing face recognition models.

Abstract

SARS-CoV-2 has presented direct and indirect challenges to the scientific community. One of the most prominent indirect challenges advents from the mandatory use of face masks in a large number of countries. Face recognition methods struggle to perform identity verification with similar accuracy on masked and unmasked individuals. It has been shown that the performance of these methods drops considerably in the presence of face masks, especially if the reference image is unmasked. We propose FocusFace, a multi-task architecture that uses contrastive learning to be able to accurately perform masked face recognition. The proposed architecture is designed to be trained from scratch or to work on top of state-of-the-art face recognition methods without sacrificing the capabilities of a existing models in conventional face recognition tasks. We also explore different approaches to design the contrastive learning module. Results are presented in terms of masked-masked (M-M) and unmasked-masked (U-M) face verification performance. For both settings, the results are on par with published methods, but for M-M specifically, the proposed method was able to outperform all the solutions that it was compared to. We further show that when using our method on top of already existing methods the training computational costs decrease significantly while retaining similar performances. The implementation and the trained models are available at GitHub.