Research of face recognition based on SVD

Through the analysis of the singular value vector and left-right orthogonal characteristic matrix, this paper affirms that the singular vectors from imaging matrix decomposition can show image gray, which is related to gray distribution density; the maximum component of the singular vectors presents the position of image levels, and other components of the singular vectors show the width of image levers. Moreover, combination the maximum singular value and the other components of the singular vectors are exactly able to ascertain image gray. The decomposed left-right orthogonal characteristic matrix can show structure information of picture contour. Finally, a kind of recognition algorithm based on basis space is brought forward, emulated in the ORL and ORL_IC database. The conclusion affirms the validity of the analysis.