ISFB-GAN: Interpretable semantic face beautification with generative adversarial network

Tianhao Peng, Mu Li, Fangmei Chen, Yong Xu, Yuan Xie, Yahan Sun, David Zhang

Abstract: Highlights•A dual-branch style encoder is proposed to obtain interpretable facial semantics.•A attention-guided generator is designed to preserve the background information.•A hybrid loss function is proposed to make facial style semantics disengaged and to enhance the learning ability of the model.