Recent Advances in Deep Learning Techniques for Face Recognition

Abstract

In recent years, researchers have proposed many deep learning (DL) methods\nfor various tasks, and particularly face recognition (FR) made an enormous leap\nusing these techniques. Deep FR systems benefit from the hierarchical\narchitecture of the DL methods to learn discriminative face representation.\nTherefore, DL techniques significantly improve state-of-the-art performance on\nFR systems and encourage diverse and efficient real-world applications. In this\npaper, we present a comprehensive analysis of various FR systems that leverage\nthe different types of DL techniques, and for the study, we summarize 168\nrecent contributions from this area. We discuss the papers related to different\nalgorithms, architectures, loss functions, activation functions, datasets,\nchallenges, improvement ideas, current and future trends of DL-based FR\nsystems. We provide a detailed discussion of various DL methods to understand\nthe current state-of-the-art, and then we discuss various activation and loss\nfunctions for the methods. Additionally, we summarize different datasets used\nwidely for FR tasks and discuss challenges related to illumination, expression,\npose variations, and occlusion. Finally, we discuss improvement ideas, current\nand future trends of FR tasks.\n