Gender classification with Local Zernike Moments and local binary patterns

Abstract

This study provides a new feature extraction method to gender classification. Local Zernike Moments is a method used for face recognition and proved that it is more successful than Gabor or LBP representations. In this study, LZM method is used for gender classification on FERET and LFW databases and demonstrated that it is more successful than LBP method on both databases. In the light of analysis done on the test results of these two methods, a new hybrid feature method built by combining LZM and LBP features is created and the performance rates are achieved as 99.57% for FERET and 97.71% for LFW databases by using Support Vector Machines (SVM) classifier. This indicates the superiority of the proposed method over suggested methods for gender classification on both controlled environment and real-world images.