Portrait Style Transfer with Generative Adversarial Networks

Portrait style transfer is a hot and practical direction for in-depth learning. As a deep learning model, Generative Adversarial Networks (GANs) have been widely used in image style conversion. We study Generative Adversarial Networks as a solution to the portrait style transfer problem. Here, we use GANs to recognize facial features. With large training in the conversion from plain to cosmetic drawings, this algorithm can make up the plain faces better intelligently. The experimental results provide the representation of facial image features by GANs and show the ability of character transformation and operation of portrait style.