Face recognition based on the differential mode SLPP

In this paper, a novel supervised locality preserving projection method based on differential mode is proposed. We proposed the concept of within-class separation degree and between-class separation degree. The discriminant criterion of difference mode based on within-class separation degree matrix and between-class separation degree matrix is constructed. Thus, the singular problem of within-class separation degree matrix is avoided which is caused by small sample problem in the SLPP algorithm. Then we extend the algorithm to nonlinear case by using kernel technique. Finally, the experimental results based on Yale face database, FERET face database show that the proposed method is effective.