MLP Neural Network for face recognition based on Gabor Features and Dimensionality Reduction techniques

Face recognition is very used in biometric market for many reasons. Face can be detected at distance without user's implication in enrollment process. Face recognition systems are still expanded in video surveillance areas to control access. This paper presents an experimental study on some proposed face recognition approaches by building systems with different techniques for features extraction and classification. To validate comparison between proposed approaches, we use three face image databases ORL, Caltech Faces and Face94. We demonstrated in this paper that Gabor Features applied with Linear Discriminate analysis to reduce size of dataset classified with MLP Neural Network ranks top the list of proposed approaches as well as many works done in literature.