A novel dictionary based SRC for face recognition

Abstract

The sparse representation based classification (SRC) performs not very well for small sample data. A discriminative common vector dictionary based SRC is introduced in this paper to address this issue. The contribution of this paper is that the dictionary of the proposed method is constructed by the discriminative common vector per class. The common vector represents the invariant property of each class, which is helpful to improve the performance of the proposed method for small sample database. Furthermore, the new dictionary has much less atoms than the original SRC based scheme, which reduces the computational cost. The experiments implemented on ORL, AR and LFW face databases demonstrate the effectiveness of the proposed method.