METHOD AND SYSTEMS OF FACE EXPRESSION FEATURES CLASSIFICATION ROBUST TO VARIETY OF FACE IMAGE APPEAR

文件类型：PDF文档
文件大小：770K
The present invention relates to a facial expression feature classification method and a system thereof. The disclosed facial expression feature classification method performed in a facial expression feature classification system for facial expression recognition includes the steps of : generating a variation image in an expression class corresponding to a variation factor in the expression class using training face images constituted in accordance with each expression class as to a query facial image; calculating an image difference between the variation image in the expression class and the query facial image and then defining the image difference as a facial feature; and determining an expression class corresponding to the query facial image by classifying the facial feature into a specific feature class by applying sparse expression. Like the same, the facial expression feature classification method has the advantage of showing stable facial expression recognition ability even though various variations exist in the query facial image, by using the facial feature extraction and classification method removing variation factors in the expression class unrelated to the expression shown in the query facial image.(AA) Start(BB) End(S201) Query facial image inputting(S203) Generating variation image in i^th expression class(S205) Calculating image difference between variation image and query facial image in i^th expression class(S207) Applying sparse expression to facial feature using variation image in i^th expression class(S209) Blending sparse expression coefficient vector corresponding C expression classes(S211) Determining expression class through classification(S213) Outputting expression class labelCOPYRIGHT KIPO 2016