Feature extraction for palmprint recognition using kernel-PCA with modification in Gabor parameters

Abstract

Palmprint recognition method is part of the biometric system that has a significant impact on the advancement of civilization, especially in the areas of sensing identity the person. To get the reliable system, the selection of actions to be taken include choosing a filter of skeleton method, selecting the scale orientation of Gabor method, and using appropriate a dimension reduction. The results show that the method of kernel fisher analysis (KFA), kernel principal component analysis (KPCA), linear discriminant analysis (LDA), and principal component analysis (PCA) became a leading candidate from dimension reduction. The research shows that the use of skeleton filter forwarded by scale orientation of Gabor by and the use of kPCA give better the equal error rate (EER) when compared with other researchers the same field.