Study On Facial Expression Recognition Based On Standard Model Features

Facial expression is an important body language. In our daily life，only7percentof information is expressed by oral language and55percent of information isexpressed by facial expression. Facial expression recognition means that the specialexpression state is analyzed and detected from the given expression image or video, thesubject’s specific emotion is ascertained automatically, and then an intelligentinteraction is achieved between human and computer. It has potential applications inmany areas such as robotics technology, image understanding, virtual-real technology,etc. The research on expression recognition has a very important practical significance,and will have considerable social and economic benefits.This paper is crossover study on applied mathematics and life science, studiesmainly facial expression recognition based on cortex-like mechanisms. Firstly therelevant knowledge of facial expression recognition and bio-vision are introduced, andthen a facial expression recognition algorithm based on the improved Standard ModelFeature (ISMF) is presented. Based on the biologically-inspired knowledge, ISMFmodels the biological visual nervous system process in visual perception task. TheISMF combined with support vector machine (SVM) is used for the extraction ofexpression features and facial expression classification. On JAFFE and TFEIDdatabases, experiments get the expected ideal results. The average recognition rate hasreached over85percent, and reached95percent or more if two images each person areused for training. In short, the ISMF facial expression feature based on the visualperception mechanism has very good expression characterization ability; the methodbased on ISMF-SVM has higher accuracy and stronger robustness compared withmethods at present.