Sparse regularization discriminant analysis for face recognition

Fei Yin, Licheng Jiao, Fanhua Shang, Lin Xiong, Xiaodong Wang

Abstract: Highlights•SRDA simultaneously seeks an efficient discriminating subspace and preserves the sparse representation structure.•Based on the special concatenated dictionary constructed via class-wise PCA decompositions, SRDA can learn sparse coefficient vector quickly.•In SRDA, the number of features is not limited to K−1. SRDA can effectively avoid the SSS problem.•SRDA uses label information to learn a more discriminating sparse representation structure.•The extensive and promising experimental results on four publicly available face datasets are presented.