Robust face recognition based on non-negative sparse discriminative low-rank representation
Xielian Hou, Caikou Chen, Shengwei Zhou, Jingshan Li · 2018
Due to occlusion or camouflage existing in the current face images, previous face recognition algorithms such as sparse representation classification algorithm do not take face damage into consideration during the training period, and therefore their testing performance will be degraded. In this paper, we propose a novel non-negative sparse discriminative low-rank representation algorithm (NSDLRR). First, we seek a sparse, low-rank and non-negative matrix in training samples. Then, we add a structural inconsistency constraint on this basis, make different kinds of samples as independent as possible, thereby increasing the extra recognition ability. Finally, the test samples are classified by sparse linear representation. Experimental results on different face database show that the algorithm has better performance.