Principal component analysis based cascade neural network for face recognition

A face recognition system which combines the powerful feature extraction property of principal component analysis (PCA) and classification capability of cascaded neural network is proposed in this paper. For a given data base the features are extracted using PCA. The feature set is divided into training and testing data. The training data is used to train a cascade neural network (CASNN). Testing data are used for performance of the system. This paper uses UMIST face data base. The performance is compared with more popular feed forward neural network (FFNN). The results obtained prove the efficacy of the proposed cascade Neural Network based classifier as compared to the Feed forward neural network classifier.