Real-Time Face Recognition System Using KPCA, LBP and Support Vector Machine

With increasing security threats, Biometric systems have importance in different fields. This appears clearly exactly after the rapid development that happened in power of computing. In this paper, the Design and implementation of a real-time face recognition system are presented. In such a system, Kernel principal component analysis (KPCA) and Local binary pattern (LBP) are used as feature extraction methods with the aid of support vector machine (SVM) to work as a classifier. A comparison between traditional feature extraction methods as (PCA and LDA) and a proposal methods are performed as well as a comparison between support vector neural network and artificial neural network classifier are also implemented. Two types of experiments, On-line, and Off-line experiments are done. In the On-line experiment, a new database is created and used. While in the off-line experiment, two types of databases (ORL and YALE) are used to estimate the performance and efficiency of the system. The combinations of these methods together enhances the experimental results in compare with other methods.