Research of Facial Expression Recognition Based on ASM Model and RS-SVM

Abstract

The accurate and efficient facial expression recognition method was researched in this paper. The active shape model (ASM) was a usual method of pattern recognition such as the facial features localization and facial expression recognition etc. In this paper, the support vector machine (SVM) was researched for the facial expression recognition, on the basis of the research SVM theory, thus a new and improved facial expression recognition method was proposed based on ASM and rough set SVM (RS-SVM). Firstly, the ASM algorithm was used for the facial features location, the features were extracted effectively. Moreover, the attribute reduction algorithm in the theory of rough set was introduced for the selection and classification of the extracted feature, the invalid and redundant features were filtered, finally, the classification was implemented with SVM algorithm. Simulation and experiment were carried out based on the Jaffe face database. Simulation results show that the new method has higher facial expression recognition rate than the traditional SVM method, and the recognition efficiency is improved greatly, it shows prospective application value in the pattern recognition and face recognition fields.