Age Progression and Regression using cycleGAN

Face detection and face related processing are emerging areas of research in computer vision and are rapidly growing topics of large interest. Facial age progression and regression is one of such interesting new generation applications. Face processing techniques have been developed during the last decade to address the difficulty of producing realistic aged faces for various smart system applications. In this paper, age progression and regression of frontal faces is achieved using Cycle Generative Adversarial Neural Network (CycleGAN). CycleGAN is a variant of the GAN architecture which is used to generate images with effects of face aging and rejuvenation. UTKFace dataset, which is an unpaired face dataset, is used for training of the proposed network. The trained network is evaluated subjectively as well as objectively for the accuracy of the age progression and regression of various face Images.