Single-Sample Face Recognition Based on WSSRC and Expanding Sample

This paper proposes a face recognition method with one training image per person, and it is based on compressed sensing. We apply nonlinear dimensionality reduction through locally linear embedding and sparse coefficients to generate multiple samples of each person. These generated samples have multi-expressions and multi-gestures are added to the original sample set for training. Then, a super sparse random projection and weighted optimization are applied to improve the SRC. This proposed method is named weighted super sparse representation classification (WSSRC) and is used for face recognition in this paper. Experiments on the well-known ORL face dataset and FERET face dataset show that WSSRC is about 15.53 % and 7.67 %, respectively, more accurate than the original SRC method in the context of single sample face recognition problem. In addition, extensive experimental results reported in this paper show that WSSRC also achieve higher recognition rates than RSRC, SSRC DMMA, and DCT-based DMMA.