Unsupervised feature extraction based on kernel discriminant projection analysis for face recognition

This paper presents a novel face recognition method based on the unsupervised discriminant projection and using a kernel based algorithm, for discriminating purposes, namely complete kernel unsupervised discriminant projection(CKUDP). This nonlinear reduction dimension algorithm using kernel function, it handles nonlinearity efficiently. Moreover, a complete solution for obtaining the optimal feature vectors in feature space is presented which can preserve the discriminant information. Experiments on the ORL database validate that by using three different methods. Experiments show that consistent and promising results are obtained.