Double direction matrix based sparse representation for face recognition

Abstract

Robust sparse representation is a well-known method in computer vision. Several sparse representation models have been proposed and perform well in face recognition. Most of them use transformed images of one dimensional vector, and such implementation ignores structural information between features. To make use of this structural information, this paper presents a novel model for face recognition, called double direction L2,1-norm based sparse representation. Unlike traditional sparse regression measuring differences between test sample and predicted response by vector norm, our model uses matrix norm, L2,1, to calculate residual. Instead of treating each pixel independently, the residual of a pixel is effected by all others in the same line and the same column by means of double direction L2,1-norm. And then, we use the alternating direction method of multipliers approach to optimize proposed model. Just as the L2,1-norm concerns, experiments show that our proposed method is more robust than other sparse methods.