Variable Lighting Face Recognition Technology

As one of the most challenging problems in pattern recognition and machine vision, face recognition has a wide range of potential applications in the areas of public security, information security and human-computer interaction. After more than 30 years, face recognition technology under well-controlled environment is going to practical stage, while the recognition performance dramatically degrade under uncontrolled environment such as variant illumination, head poses, facial expressions, occlusion on face and so on. Therefore, one of key issues in face recognition is the study of robust recognition algorithm. In this paper, our research focuses on face recognition with illumination variation. The major tasks done by the paper are summarized as follows:Firstly, in the part of image preprocessing, a novel approach of illumination compensation is proposed to handle this problem in this paper. First, a binary image is constructed from the original face image, and it can be used to determine the direction of light. And then, the original image is compensated with one average intensity difference which corresponds to the same direction of light. Finally, the face recognition is done without first three eigenfaces. Experimental results show that this method can increase face recognition rates under different illuminations. Secondly, in the part of feature extraction, four methods are introduced: PCA, 2D-PCA, 2D-LDA, PCA+LDA. In this four methods, however, 2D-PCA and 2D-LDA are both based on 2D image matrices rather than 1D vector, so the image matrix does not need to be transformed into a vector prior to feature extraction.Then in the part of classification recognition, it also introduces four typical classification recognitions: Nearest neighbor classification, K-nearest neighbor classification, SVM and Bayesian classification. And it introduces the Bayesian basic method and study the method applied to the face recognition in detail.Finally, in the experiment, different methods of feature extraction are brought forward to associate different classification with the raw images, the Normal-Histogram images and the compensated images respectively. Then from analyzing the experimental data, a good association is obtained to get good result under variant illumination.