Face Recognition Using Matrix Decomposition Technique Eigenvectors and SVD

Principle Component Analysis (PCA) is an important and well-known technique of face recognition, where eigenvectors are used. In this paper, we propose a face recognition technique, which combines Eigenvectors with Singular Value Decomposition (SVD) techniques to reduce size of the Eigen-matrix. The detailed theoretical derivation and analysis are presented and a simulation results on Olivetti Research Laboratory (ORL) face database has given. The simulation result indicates that the proposed approach is superior to conventional PCA with lower database size and recognition performance.