Real-time training method of expression database, and feedback mechanism of expression database

Abstract

The invention discloses a real-time training method for an expression database. The method comprises the following steps: S1, inputting in advance: enabling an actor to input expression coordinate sequences of all expressions of the expression database in advance before performance; S2, presetting a range: obtaining a preset range through big data analysis of values of the multidirectional tensionsensors corresponding to all facial feature point positions of each expression; S3, carrying out expression real-time acquisition: when an actor performs, making different expressions, and generatinga set of expression coordinate sequence consistent with the positions of the facial feature points in step S1; S4, comparing and screening: comparing and screening the numerical values of the multidirectional tension sensors corresponding to the positions of the facial feature points of the expression coordinate sequence; and S5, performing multi-sample training: continuously and accurately determining a predetermined range of the facial feature points through a large number of expression coordinate sequences. The motion state of the facial muscles is captured in real time in the performanceprocess, so that the preset range of each muscle of each expression of a specific actor is continuously updated and accurately determined, and the purpose of training a database is achieved.

Description

technical field [0001] The invention relates to the technical field of database updating, in particular to a real-time training method of an expression database and a feedback mechanism of the expression database. Background technique [0002] Facial expression recognition has become a hot topic in recent years. Emoticon is the most effective way in human emotional communication, which contains a lot of personal behavior information. Human facial expressions usually include happiness, sadness, anger, fear, surprise, and disgust. By establishing a facial expression database, it can provide convenience for post-production film and television special effects and character facial animation, and reduce the cost of post-production image processing. [0003] The traditional facial expression database is a facial expression recognition algorithm based on neural networks, which requires a large number of training pictures to form an expression database. Training pictures are usual...

AI Technical Summary

Problems solved by technology

Training pictures are usually collected manually, so the scale of pictures is largely limited, resulting in insufficient number of training samples

This kind of facial expression database focuses on the study of facial expression features in static images, but the overall features cannot describe the detailed information of human faces, and it is easy to lose the detailed features of expressive faces, so that the trained features lose the detailed information of various expressions. Poor facial expression recognition

Images