Gender classification by facial feature extraction using topographic independent component analysis

Abstract

Recognition of gender from face image has attracted a huge attention now a days. Many identification systems are being developed to identify a person, as most of the technique for gender classification stand on facial features. In this paper, we presented a gender classification framework consist of a series of phases for determining the gender as the final output. Initially we start by detecting the face from an image using Viola Jones and then extract the facial feature using the Topographic Independent Component Analysis. The features extracted here are used to train the SVM classifier for the classification step. Our experimental result gives the best accuracy in determining the images as of male or female and gives average performance of 96% correct gender identification on images.