Relaxed collaborative representation for face recognition based low-rank matrix recovery

Face recognition is of paramount importance in computer vision and biometrics systems. In this paper we propose an improved method which is suitable to handle variations in image configurations like pose, illumination, and facial expressions as well as occlusion and disguise, in order to provide high efficiencyi in the face recognition. This method integrates the low-rank matrix which is recovered by using robust principal component analysis (RPCA) with relaxed collaborative representation (RCR). Low-rank representation allows us to better discriminate information which benefits to face identification, and R-CR contributes to the reduction of the variance of coding vector after coding each feature vector on its associated dictionary to allow flexibility of feature coding, thus addressing the similarity among features. Furthermore, it is characterized by the exploitation of the distinctiveness of different features by weighting its distance to other features in the coding domain. The effectiveness of the proposed method is validated by extensive experiments on different benchmark face databases.