期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2014
卷号：2
期号：7
出版社：S&S Publications
摘要：This project entitled “Comparative analysis of advanced Face Recognition Techniques”, it is basedon fuzzy c means clustering and associated sub neural network. It deals with the face is a complex multidimensionalvisual model and developing a computational model for face recognition is difficult. In this paper, it represents amethod for face recognition base on similar neural networks. Neural networks (NNs) have been widely used invarious fields. However, the computing effectiveness decreases rapidly if the scale of the NN increases. In thispaper, a new method of face recognition based on fuzzy clustering and parallel NNs is proposed. The face patternsare divided into several small-scale neural networks based on fuzzy clustering and they are combine to obtain therecognition result. The facial feature vector was compared by PCA and LDA methods. In particular, the proposedmethod achieved 98.75% recognition accuracy for 240 patterns of 20 registrants and a 99.58% rejection rate for 240patterns of 20 nonregistrants. Experimental results show that the performance of our new face-recognition method isbetter than those of the LDA based face recognition system.
