A Face Recognition Scheme Based on Weighted Face Features and Wavelet Transform

Abstract

In face recognition, face feature extraction and reduction of the dimensions is the key question to face recognition. The face recognition algorithm is based on wavelet transform only in the low-frequency component of the wavelet decomposition to extract image features for classification. The results have a lot of parts of the high-frequency components to identify favorable information. In order to extract more effectively facial image features, the proposed face recognition algorithm is based on wavelet transform and feature weighting fusion. Firstly, wavelet transform can reduce the dimension of the human face images. Four wavelet sub-bands have respectively extract features using principal component analysis (PCA) and feature weighted fusion. Finally, the use of support vector machine (SVM) classification is important. The experiments on the ORL database and the recognition accurate rate can reach 98.50%. The experiment results are shown that the algorithm can effectively improve the face recognition ability, it is compared with traditional recognition algorithm has high recognition accuracy and recognition speed.