Face Recognition System using Discriminant Analysis on Support Vectors

摘要

The advantages of Principal Component Analysis (PCA) for feature extraction, Support Vector Machines (SVMs) algorithm for pattern recognition and Linear Discriminant Analysis (LDA) recognition technique are combined to reduce the error rate in face recognition. A hybrid technique (PCA+SV+LDA) using PCA along with support vectors and LDA is presented in this paper. Large sample size problem of LDA is reduced to small sample size problem using support vectors, and discriminant analysis is done only on support vectors. Experiments are performed on Indian Face database and AT&T Face database and error rates of classification and elapsed time for performance evaluation are compared with techniques such LDA, PCA, PCA+LDA and PCA+SVM.