Improving Variational Autoencoder with Deep Feature Consistent and Generative Adversarial Training

Xianxu Hou, Guoping Qiu

Abstract: We present a new method for improving the performances of variational autoencoder (VAE). In addition to enforcing the deep feature consistency principle thus ensuring the VAE output and its corresponding input images to have similar deep features, we also implement an adversarial generative training mechanism to force the VAE to output realistic and natural images. We present experimental results to show that the VAE trained with our new method outperform state of the art in generating face images with much clearer and more natural noses, eyes, teeth, hair textures as well as reasonable backgrounds. We also show that the VAE trained with the new method can extract more effective features that outperform state of the art in facial attribute recognition.

Keywords: Variational autoencoder, VAE, generative adversarial network, GAN, Facial Attribute Recognition