A survey on video face recognition using deep learning

Abstract

The research on facial recognition consists of Still-Image Face Recognition (SIFR) and Video Face Recognition (VFR), is a common subject being debated among researchers since it does not require any touch like other biometric identification, such as fingerprints and palm prints. Various methods have been proposed and developed to solve the problems of face recognition. Convolutional Neural Network (CNN) is one of the deep learning techniques that is suggested for both SIFR and VFR. However, several issues related to VFR have still not been solved. Hence, the objective of this paper is to review VFR using deep learning that specifically focuses on several steps of VFR. The VFR steps consists of six main stages; input video of the face, face anti-spoofing module, face and landmark detection, preprocessing, facial feature extraction and face output that include identification or verification result. A summary of implementation of deep learning within VFR steps is discussed. Finally, some directions for future research are also discussed.