Patch-based Sparse Dictionary Representation for Face Recognition with Single Sample per Person

In this paper, we solve the problem of robust face recognition (FR) with single sample per person (SSPP). FR with SSPP is a very challenging task due to in such a scenario lacking of information to predict the variations of the query sample. We propose a novel method patch-based sparse dictionary representation (PSDR) to tackle the problem of various variations e.g. expressions, illuminations, corruption, occlusion and disguises in FR with SSPP. The key idea of our scheme is to combine a local sparse representation and a patch-based generic variation dictionary learning to predict the possible facial variations of query image and classification. To extract more feature information in classification, we adopt a patch-based method. Our experiments on Extended Yale B and AR databases show that our method outperforms the state-of-art approaches.