LNMF Learning for color face recognition

In this paper, a novel face recognition method named local NMF learning (LNMF Learning) is proposed. Color face is first decomposed into R, G and B components. Component data of the same color channel is aligned together and encoded through matrix mode respectively. Local nonnegative matrix factorization (LNMF) method is employed to compute facial features. Projective coefficients on base images are utilized as features for recognition task. Experimental results on CVL and CMU PIE face databases verify the effectiveness of the proposed approach.