Solving Age-Invariant Face Recognition and Face Age Synthesis simultaneously using Deep Learning

Abstract

Face recognition has been a hot research topic in computer vision for many years. Recently, deep learning-based methods achieve excellent performance, even surpassing human in several scenarios by empowering the face recognition models with deep learning networks. Despite the remarkable success of general face recognition, how to minimize the effects of age variation is a challenge for current face recognition systems to correctly identify faces in many practical applications such as finding loss children. This project will introduce a unified, multi-task framework to jointly handle two tasks, termed MTLFace, which can learn age-invariant identity-related representation while achieving pleasing face synthesis