Pattern Recognition by Means of Linear Discriminant Analysis and the Principal Components Analysis

摘要

This paper is devoted to the problem of pattern recognition solved by methods of principal com- ponents and linear discriminant analysis. The efficiency of the described method is studied for a case when pictures of faces have not yet undergone preliminary processing that would have led them to the standard form (scale, centering, background clipping, brightness adjustment). When processing large sets of images in order to reduce the complexity of computation of principal components, it is proposed to use the linear condensa- tion method and principal component synthesis. We have studied the effectiveness of the approach based on principal component analysis and linear discriminant analysis using the linear condensation method and principal component synthesis on the ORL and FERET database of face images.