Statistics Based Automatic Face Recognition System

The automatic recognition of human faces is an active subject in the area of computer vision and pattern recognition over the past few years, which has a wide range of potential applications in the areas of public security, identification of certificate, entrance control and video surveillance. This paper is a study of the recognition of pose or expression varied human faces image. Firstly, human facial cognitive models and rules are studied from the view of cognitive psychology. Then the traditional method forface recognition-eigenface is analyzed. A improved approach-based-symmetrical half-PCA (SHPCA) is introduced for recognizing pose and expression varied human faces. The essential idea of the approach is, according to the idea of half eigenface, to segment the face image into the upper and lower part for reducing the effect of different expression, then project the mirror image of face image into the respective space of half eigenface for enhancing the ability of pose varied human face recognition. In the part of recognition, the projection vectors with different weight are applied to compute the Euclidean discriminate distance and then discriminate the face image with the rule of the least distance. The preprocessing techniques for human face images are used to reduce the influences of illumination, location and scale variation of faces. 400 pose and expression varied faces of 40 persons from ORL face database that is used widely by international researcher are tested in experiments. Eigenface gains 93.5% correct recognition rate, whereas this approach gains 97.5% with five faces refused mistakenly for exaggerated expressions. Experiments results show that this method gains better recognition rate than eigenface, especially for pose and expression varied human faces.