m fathi - iran uinversity of science and technology
gholamreza ramezan

principal component analysis method is mor considered among the algebraic methods for extracting face features. This analysis is a strong statistical method in thid paper the performance of this method on the ORL olivetti research library face images bank is presented the tests we performed has shown the suitable threshold in PCA matrix this threshold shows the necessary lines of PCA matrix for a face recognition systems our researches have shown that the suitable threshold in 150 eigenvectors and in this case the recognition rate is 84% .

face recognition , principal component analysis PCA, wavelet