A fuzzy neuro clustering based vector quantization for face recognition

Abstract

A face recognition system is a computer application for automatically identifying or verifying a person from a digital image or a video frame. In this paper, an improved codebook design method is proposed for Vector Quantization (VQ)-based face recognition which improves recognition accuracy. A codebook is created by combining a systematically organized codebook based on the classification of code patterns and another codebook created by Integrated Adaptive Fuzzy Clustering (IAFC) method. IAFC is a fuzzy neural network which incorporates a fuzzy learning rule into a neural network. The performance of proposed algorithm is demonstrated by using publicly available AT&T database and Yale database. Experimental results show face recognition using the proposed codebook is more efficient yielding a rate of 99.25% for AT & T and 98.18% for Yale which is higher than most of the existing methods. © 2011 Springer-Verlag.