Face Recognition with Discrete Cosine Transform

Face recognition has been a hot issue of concern to people, existing algorithms mainly for how to improve face recognition accuracy, and reduce the use of the feature vector dimension. A face recognition method based on the discrete cosine transform and SVM-based feature classification is presented. Face Recognition on the DCT coefficients as feature vectors in the problem of effective choice. In the training process of SVM, the training sample is divided into two parts, one part of the training samples using the test results of feature selection, and use the resulting accuracy rate as the feature selection criteria, the proposed algorithm on ORL face image standard library of verified experiments, and achieved good results prove the effective and reliable.