Real time face identification using a neural network approach

Abstract

This paper presents techniques for developing a system for identification of faces in real time based on biometric technology [22], where the identification phase is being implemented by an artificial neural network. The motivation for this research work stems from the observation that the human face provides a particularly interesting structure. Face images are obtained by a web camera and then used for the digital image preprocessing techniques. Feature extraction techniques are applied; the extracted image features are fed to the neural network for learning. Due to the fact that the effectiveness of systems for identification techniques rely primarily on Preprocessing and feature extraction, therefore, this work presents different features extraction techniques, and a comparison between methods is made, in terms of their percentages of recognition. We described the most used techniques for this task [10], i.e.: Edge extraction, Wavelet Analysis, eigenfaces. © 2010 Springer-Verlag Berlin Heidelberg.