A survey on disguise face recognition

Abstract

ABSTRACT Human face is one of the physiological traits used in biometric system and gained a significant attention in recent years due to the advancement in technology. It has been remarkable progress in face recognition system since 1960 owing to advances in 3D-face modeling and analysis, face detection, and tracking. The reason for this progress is to provide security in surveillance systems, criminal identification, border control, security access, and many more. Despite this, proposed techniques are not robust enough for many challenging issues like variations in pose, illumination, expression, and disguise. This paper provides a survey on some of the well-known deep learning methodologies focused on these issues of face recognition. Furthermore, the performance of these methodologies is also studied in terms of accuracy, computational cost, and robustness. In addition to this, it also summarizes and outlines the importance of disguise face datasets. Finally, open challenges involved in deep learning framework, disguise face dataset, and the impact of disguise co-variate on the current face recognition system are discussed for future research.