Robust Face Recognition Based on Supervised Sparse Representation

Abstract

Sparse representation-based classification (SRC) has become a popular methodology in face recognition in recent years. One widely used manner is to enforce minimum $l_{ 1}$-norm on coding coefficient vector, which requires high computational cost. On the other hand, supervised sparse representation-based method (SSR) realizes sparse representation classification with higher efficiency by representing a probe using multiple phases. Nevertheless, since previous SSR methods only deal with Gaussian noise, they cannot satisfy empirical robust face recognition application. In this paper, we propose a robust supervised sparse representation (RSSR) model, which uses a two-phase scheme of robust representation to compute a sparse coding vector. To solve the model of RSSR efficiently, an algorithm based on iterative reweighting is proposed. We compare the RSSR with other state-of-the-art methods and the experimental results demonstrate that RSSR obtains competitive performance.