Face Recognition Using a Fuzzy-Gaussian Neural Network Victor-Emil NEAGOE and Iuliana-Florentina IATAN Depart. of Applied Electronics and Information Eng., POLITEHNICA University of Bucharest, Bucharest, 77206 Romania Email: vneagoe@xnet.ro Abstract We present a face recognition approach using a new version of Chen and Teng fuzzy neural network [1], which we have modified from an identifier into a neuro- fuzzy classifier called Fuzzy-Gaussian Neural Network (FGNN). We have deduced the modified equations for training the FGNN. Our presented face recognition cascade has two stages: (a) Feature extraction using either the Principal Component Analysis (PCA) or the Discrete Cosine Transform (DCT); (b) Pattern classification using the FGNN. We have performed the software implementation of the algorithm and have experimented the face recognition task for a database of 100 images (10 classes) . The recognition score has been of 100% (for the test lot) for almost all the considered variants of feature extraction .We have also compared the performances of the FGNN with those obtained using a classical multilayer fuzzy perceptron (FP). We can deduce the significant advantage of the proposed FGNN over FP. 1. Introduction Face recognition has been studied for many years and has practical and/or potential applications in areas such as security systems, criminal identification, video telephony, medicine, and so on. In comparison to other identification techniques, face recognition has the advantage of being non-intrusive and requiring very little cooperation; it has become a hot topic recently as better hardware and better software have become available. On the other side, the hybrid systems of fuzzy logic and neural networks (often referred as fuzzy neural networks) represent exciting models of computational intelligence with direct applications in pattern recognition, approximation, and control. We further consider a modified version of the fuzzy neural network described by Chen and Teng [1], used as identifier in control systems, by transforming this network from an identifier into a performing classifier that we called Fuzzy Gaussian Neural Network (FGNN). We have applied this model in a face recognition cascade having the following processing stages: (a) feature extraction using either Principal Component Analysis (PCA) or Discrete Cosine Transform (DCT); (b) FGNN as a classifier. The results of computer simulation are given. 2. Fuzzy Gaussian Neural Network (FGNN) 2.1. Architecture The four-layer structure of the Fuzzy-Gaussian Neural Network (FGNN) is shown in Fig. 1. It represents a modified version of Chen and Teng fuzzy neural network, by transforming the function of approximation into a function of classification. The change affects only the equations of the fourth layer, but the structure diagram is similar. Its construction is based on fuzzy rules of the form j ℜ : If 1 x is j 1 A and 2 x is j 2 A … and m x is j m A , then 1 y is j 1 β , …, M y is j M β , where m is the dimension of the input vectors (number of retained features), and j is the rule index (j=1,…, K). The number of output neurons (of the fourth layer) corresponds to the number of classes and it is equal to M. The j-th fuzzy rule is illustrated in Fig. 2. The FGNN keeps the advantages of the original fuzzy net described by Chen and Teng [1] for identification in control systems: (a) its structure allows us to construct the fuzzy system rule by rule; (b) if the prior knowledge of an expert is available, then we can directly add some rule nodes and term nodes; (c) the number of rules do not increase exponentially with the number of inputs; (d) elimination of redundant nodes rule by rule. Proceedings of the First IEEE International Conference on Cognitive Informatics (ICCI’02) 0-7695-1724-2/02 $17.00 © 2002 IEEE