Biometrics in the Time of Pandemic: 40% Masked Face Recognition Degradation can be Reduced to 2%

TL;DR

This paper addresses the challenge of face recognition with masked faces during pandemics, demonstrating a significant reduction in performance degradation from over 36% to less than 2% using advanced deep learning methods.

Contribution

The study introduces a deep learning approach that substantially reduces masked face recognition degradation in cross-spectral scenarios, improving accuracy during pandemics.

Findings

Recognition degradation reduced from 36.78% to 1.79%.

Deep learning approaches outperform traditional methods.

Effective in border checkpoint scenarios.

Abstract

In this study of the face recognition on masked versus unmasked faces generated using Flickr-Faces-HQ and SpeakingFaces datasets, we report 36.78% degradation of recognition performance caused by the mask-wearing at the time of pandemics, in particular, in border checkpoint scenarios. We have achieved better performance and reduced the degradation to 1.79% using advanced deep learning approaches in the cross-spectral domain.