Application of Non-negative sparse matrix factorization in occluded face recognition

In order to reduce the impact of block for the rate of face recognition ,in this paper, through the control of sparseness in the non-negative matrix factorization , the face image do non-negative sparse coding to obtain the eigenspace for the image. The experiment uses the ORL face database. The experimental results show that using NMFs obtains Eigenfaces with the local features of face and has a strong ability to express the occluded human face. The algorithm has good adaptability to partial occlusion, and has better robustness than PCA algorithm.