Semantic Facial Image Manipulation is a conditional generation task. It opens the door to many exciting applications in diﬀerent areas, including the movie industry, photography technologies, fashion and e-commerce business. The goal is to synthesize a facial image conditioned on both the identity and expression.
This project aims at generating 2D photo-realistic facial images with accurate target expression labeled in FACS coding guided by 3D information. It can be applied to tasks like expression recognition and FACS data generation.