A recognition method and system for a facial action unit

Abstract

The invention relates to an identification method and system of a facial action unit. According to the invention, a facial action unit recognition model is contructed based on historical target facialfeature points, historical target spatial forms, historical spatial feature characterization parameters, characterization index change ranges of the historical spatial feature characterization parameters and weight relations of the historical spatial feature characterization parameters, and then facial action unit motion recognition is carried out. The recognition of the facial action unit can recognize fine changes of facial expressions, and results can be used for analysis of facial micro-expressions and analysis of facial expressions. The method provided by the invention has the advantagesof low requirements on sample library sample quality and quantity, high calculation speed, short parameter adjustment time, simplicity in operation, no need of a large amount of cost input, capability of ensuring high-efficiency accuracy, high recognition capability on a target facial action unit (AU), recognition capability no matter whether the target facial action unit (AU) occurs independently or in a combination manner, and simplicity in application.

Description

technical field [0001] The invention relates to recognition of facial action units, in particular to a recognition method and system for facial action units. Background technique [0002] The Facial Action Coding System (FACS), based on a system originally developed by the Swedish anatomist Carl-Herman Hjortsjlian, classifies human facial movements by facial appearance. It was later adopted by Paul Ekman and WallaceV.Friesen, and made in-depth research. Through observation and biofeedback, it described the corresponding relationship between different facial muscle movements and different expressions, and published it in 1978. [0003] The facial action coding system (FACS), according to the anatomical characteristics of the human face, can be divided into a number of action units (Action Units, AU) that are both independent and interrelated. The movement characteristics of these facial action units and the functions they control The main area can reflect facial expressions....

AI Technical Summary

Problems solved by technology

However, the existing AU recognition algorithms rarely have the setting of this link, which will lead to the introduction of recognition errors in the early stage of recognition, and further lead to any subsequent work based on facial action units (AU), such as expression recognition, micro-expression recognition, Intensification of identification errors in status identification and other links

Images