Facial expression synthesis method based on geometric prior adversarial generative adversarial network

• Main Authors: SUN ZHENAN, HE RAN, HOU LUANXUAN, MA XIN
• Format: Patent
• Language: Chinese; English
• Published: 09.04.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses a facial expression synthesis method based on a geometric prior adversarial generative network. The method comprises the following steps: carrying out the preprocessing of the image data of an expression database, extracting the expression key points of a face image, and making a face key point thermodynamic diagram; training two pairs of generators and discriminators in the adversarial generative network by taking the human face image and the human face key point thermodynamic diagram as the input of the network model to complete the tasks of expression generation and removal at the same time; and training to obtain a geometric priori adversarial generative network model capable of simultaneously performing expression generation and expression removal, performing expression generation and removal processing on the test data by using the trained geometric priori adversarial generative network model, and performing expression-invariant face recognition on the image subjected to the remo