Weighted Group Sparse Representation Based on Robust Regression for Face Recognition

Recently, sparse representations have attracted a lot of attention. In this paper, we present a novel group sparse representation based on robust regression approach (GSRR) by modeling the sparse coding as group sparse constrained robust regression problem. Unlike traditional group sparse representation, we propose a weighted group sparse penalty which integrates similarity between the test sample and distinct classes and data locality. An efficient iteratively reweighted sparse coding algorithm is proposed to solve the GSRR model. The proposed classification algorithm has been evaluated on three publicly available face databases under varying illuminations and poses. The experimental results demonstrate that the performance of our algorithm is better than that of the state of the art methods.