The Technique Research Of Multi-Minorities Facial Expression Understanding And Analysis

Expression is an important non-verbal behavior of human beings to express their emotional information and the important reaction of the inner life of man, thus to correctly identify the facial expression is a precondition to ensure the normal communication. Due to the difference of genetic, areas and culture, different ethnic groups in China have a certain difference in the facial expression when express their emotions. This paper put forward the idea of study the angles which are consist of different facial feature points and using Facial Action Units (AU) which is based on Facial Action Coding System (FACS) and facial feature parts to understand and analyze Chinese multi-minorities facial expression images, and then study the differences and the inherent laws of different ethnic groups when they express their facial expression.In this paper, using active shape model (ASM) fitting CANDIDE-3model to locate multi-minorities facial expression images’feature points, it can not only accurately locate facial feature points, but also can accurately extract facial feature parts and facial action units (AU). In this paper mainly studies facial expression images of Mongolian, Uygur, Zhuang, Tibetan, and compares their expression images with neutral facial expression images. By the study of the motion of displacement of feature points that control facial feature components such as eyebrows, eyes, nose, mouth, and through machine learning can draw different table image to integrate and analyze facial expression information of different ethnic groups. In this paper also put forward the idea that by the study of the angles consist of different feature points make the understanding and analysis of multi-minorities facial expression come true. At the same time gives the test results of expression images. The method put forward in this paper provides a new idea for the research of Chinese multi-minorities facial expression.The purpose of this thesis is to provide a new method for the research of Chinese multi-minorities facial expressions understanding, recognition and analysis and the feature law. In addition, the research of the laws of different minorities’facial expression can be used for the preservation and heritage of ethnic culture, and the related results can be widely used in the national digital cultural industries.