Classifying gender from faces using independent components

Abstract

Gender classification algorithms uses a number of dimensionality reduction techniques. These techniques are used to remove linear and non-linear dependency among the data. Most popular techniques are Principal Component Analysis (PCA) and Independent Component analysis (ICA). PCA makes the data uncorrelated while ICA gives independent data. In this approach, infomax ICA has been used for gender recognition and it is outperforming PCA. To perform classification using ICA features, Support Vector Machine (SVM) has been used. For larger training set SVM is performing with an accuracy of 98%. The accuracy values are obtained for change in size of testing set and the proposed system performs with an average accuracy of 96%. © 2011 Springer-Verlag Berlin Heidelberg.