Development of a face recognition system using artificial intelligent techniques based on hybrid feature selection

Abstract

This paper introduces a new solution of recognizing human faces in 2-dimensional digital images using a localization of facial parts information (such as eyes, nose, eyebrow, mouth, ears, etc), Pseudo Zernike Moment Invariants (PZMI) as feature, and Radial Basis Function (RBF) neural network as classifier. The proposed method, employs a set of different kind of feature from the face images and projected in each appropriately transform methods in parallel. Then the output of the RBF classifiers are fused together to make a decision. The proposed method can process the recognition and lends itself to higher classification accuracy by providing flexibility in dealing facial features, and not affected by irrelevant information in an image