Facial expression synthesis method based on generative adversarial network

Abstract

The invention puts forward a facial expression synthesis method based on a generative adversarial network. The method comprises the following main contents of a geometric guidance facial expression synthesis and facial geometric operation. The process of the geometric guidance facial expression synthesis and facial geometric operation comprises the following steps that: firstly, giving a heat picture with a target facial expression and a front face without expressions to correspondingly synthesize a new face image; then, carrying out weighted summation on all loss functions to obtain a total loss function; thirdly, adopting the geometric positions of one group of datum points to guide facial expression editing, and using an expression synthesis model to obtain a facial expression transferring result; and finally, linearly regulating a shape parameter value to carry out facial expression interpolation. By use of the method, the geometric guidance generative adversarial network can be used for generating vivid images of different expressions from single image, the synthesis image is subjected to fine granularity control, facial expression transfer and interpolation can be easily carried out, and facial expression transfer and cross expression identification can be realized.

Description

technical field [0001] The invention relates to the field of facial expression synthesis, in particular to a method for synthesizing facial expressions based on generative confrontation networks. Background technique [0002] Human face plays a very important function of information expression in human communication, it conveys human emotion and mental state. In recent years, the automatic processing of facial expressions by computer has become a hot research topic in the fields of computer vision, computer graphics, and pattern recognition. It has broad application prospects in video conferencing, film and television production, and intelligent human-computer interface. Facial expression processing includes facial expression recognition and facial expression synthesis. Among them, facial expression synthesis makes it easier for people to use equipment. For example, facial expression synthesis enables computers to generate delicate and realistic facial expression animations...

AI Technical Summary

Problems solved by technology

However, although the variational autoencoders utilized by traditional methods can produce high-resolution realistic images, the calculations are very complex, while the images generated by deep generative models often lack details, and the images are blurry or low-resolution

Images