Deep Face Recognition: A Tutorial

Abstract

Face recognition made tremendous leaps in the last five years with a myriad of systems proposing novel techniques substantially backed by deep convolutional neural networks (DCNN). Though face recognition performance skyrocketed using deep-learning, leading to the belief that this technique reached human performance, yet unconstrained face recognition remains an open problem. As deep-learning allowed achieving nearly perfect accuracy on the LFW dataset, the newly released IJB sets showed how face recognition remains a difficult problem in unconstrained environments. This tutorial summarizes the main advances in deep face recognition and, more in general, in learning face representations for verification and identification. The tutorial provides a clear, structured presentation of the principal, state-of-the-art face recognition techniques appeared in the last five years in top computer vision venues. The participants will be guided firstly to understand the face recognition problem and its evaluation criteria (closed-set or open-set identification; verification) and, then, to hear about very recent methods for deep face recognition.