Face Recognition Approach by Subspace Extended Sparse Representation and Discriminative Feature Learning
Mengmeng Liao,Xiaodong Gu
DOI: https://doi.org/10.1016/j.neucom.2019.09.025
IF: 6
2019-01-01
Neurocomputing
Abstract:To address the problem of face recognition where the number of the labeled samples is insufficient and those samples involve pose, illumination and expression variations, etc., this paper proposes a face recognition approach by subspace extended sparse representation and discriminative feature learning, called SESRC & LDF. In SESRC&LDF, each test image is considered to be the image with small pose variation or the image with large pose variation according to its symmetry. For each test image, if it is considered to be the former, it will be recognized by the proposed subspace extended sparse representation classifier (SESRC), otherwise, it will be recognized by the face recognition method based on learning discriminative feature (LDF) proposed in this paper. On eight benchmark face databases, including Yale, AR, LFW, Extended Yale B, FEI, FERET, UMIST and Georgia Tech, empirical results show that SESRC & LDF achieves the highest recognition rates, outperforming many algorithms. Those algorithms include some state-of-the-art ones, such as PLR, MDFR and OPR.