Reflectance analysis based countermeasure technique to detect face mask attacks

Abstract

Face photographs, videos or masks can be used to spoof face recognition systems. Recent studies show that face recognition systems are vulnerable to these attacks. In this paper, a countermeasure technique, which analyzes the reflectance characteristics of masks and real faces, is proposed to detect mask attacks. There are limited studies on countermeasures against mask attacks. The reason for this delay is mainly due to the unavailability of public mask attack databases. In this study, a 2D+3D face mask attack database is used which is prepared for a research project in which the authors are all involved. The performance of the countermeasure is evaluated using the texture images which were captured during the acquisition of 3D scans. The results of the proposed countermeasure outperform the results of existing techniques, achieving a classification accuracy of 94.47%. In this paper, it is also proved that reflectance analysis may provide more information for the purpose of mask spoofing detection compared to texture analysis.