Face Recognition Based on Generalized Kernel Fisher Discriminant Vectors and BP Network Classifier
A new GKFDA-based face recognition system is presented in this paper. The kernel mapping is applied to project the original sample space into an implicit space F, and the procedure of extracting the LDA optimal discriminate vectors is converted to extract the kernel optimal discriminant vectors in space F. A new method based on the orthogonal complementary space is proposed to calculate the kernel optimal discriminant vectors. In addition, a BP network classifier is designed to classify. The experimental result for the ORL database shows that the proposed method is superior to other methods in terms of recognition accuracy, efficient and applicability.
Liqiang Zhao Xiaohua Zhang Kong Lingfu Zhifei Liu
Department of Mathematics & Physics, Heibei Normal University of Science & Technology, Qinhuangdao, Yanshan University, Qinhuangdao, 066004 China Division of Personnel Affairs, Heibei Normal University of Science & Technology, Qinhuangdao, Hebei,
国际会议
武汉
英文
145-150
2008-12-19（万方平台首次上网日期，不代表论文的发表时间）