Application of Boolean Kernel Function SVM in Face Recognition
Kebin Cui, Yingshuag Du · 2009
SVM based on Boolean kernel function has outstanding performance in classifying, for the problem of face recognition, recognizing strategies based on MDNF and MPDNF Boolean kernel function SVM are Proposed. Firstly, Karhunen-Loeve transform is employed to get the representation basis of face image set, secondly, the extracted characteristics is translated into 0-1 format, thirdly, SVM based Boolean kernel function are used to classify. The face recognition experiments with ORL face databases show that the proposed methods led to significantly better recognition accuracy compared with traditional PCA method and linear SVM, between the proposed methods, the one based on MPDNF Boolean kernel function get better performance.