Study on face recognition with combined of fisher algorithm and support vector machine
Lushen Wu, Peimin Wu, Fanwen Meng, Weijing Yu · 2010
According to the face recognition, the combined algorithm of Fisherfaces and one-against-rest classifiers based on support vector machine is proposed in the paper. First the wavelet transform is used to compress the image dimension and shorten the time of training. After reducing the dimension with PCA algorithm, the Fisher linear discriminative rules are adopted to extract the optimal features of face. Then the one-against-rest classifiers of SVM are built with the features of the training sample face, which we can use to recognize the face images. The experiments are implemented on ORL and Yale face databases, and the results show that the accuracy rates are respectively 97.75% and 97.80% and the average recognition time is 9.8ms, which also demonstrates that the Fisherfaces algorithm is superior to Eigenfaces one on feature extraction.