My Eyes Are Up Here: Promoting Focus on Uncovered Regions in Masked Face Recognition

TL;DR

This paper addresses masked face recognition challenges caused by mask-wearing, proposing a combined loss method that improves verification performance between masked and unmasked faces, validated on multiple datasets.

Contribution

It introduces a novel training approach combining triplet loss and MSE to enhance masked face recognition accuracy.

Findings

Significant performance improvements with the proposed method.

Effective in masked-unmasked face verification scenarios.

Validated on two different evaluation databases.

Abstract

The recent Covid-19 pandemic and the fact that wearing masks in public is now mandatory in several countries, created challenges in the use of face recognition systems (FRS). In this work, we address the challenge of masked face recognition (MFR) and focus on evaluating the verification performance in FRS when verifying masked vs unmasked faces compared to verifying only unmasked faces. We propose a methodology that combines the traditional triplet loss and the mean squared error (MSE) intending to improve the robustness of an MFR system in the masked-unmasked comparison mode. The results obtained by our proposed method show improvements in a detailed step-wise ablation study. The conducted study showed significant performance gains induced by our proposed training paradigm and modified triplet loss on two evaluation databases.