A robust sparse representation based face recognition system for smartphones

Many research works have been done in face recognition during the last years that indicates the importance of face recognition systems in many applications including identity authentication. In this paper we propose an approach for face recognition which is suitable for unconstrained image acquisition and has a low computational cost. Since in practical applications such as in smartphones, imaging conditions are not limited to existing images in the database, robustness of the recognition algorithm is very important. Here a sparse representation framework is proposed which achieves some degree of robustness. Using double sparse representation the high computational cost of sparsity-based classifications is resolved. The experimental result indicates that the proposed method outperforms other well-known algorithms in terms of robustness and it is still fast enough for real-time applications in smartphones.