Sparse representation based face recognition under varying illumination conditions

Abstract

In this paper we have developed novel technique to recognize faces across Illumination. Illumination is a condition where an image of same individual looks different due to varying lighting conditions. Recognizing faces across Illumination is proposed using Sparse Representation technique and tested using Extended Yale B database which consist of images across varying lighting conditions. Here we have considered images of 16 individuals at different lighting conditions. From our analysis we have found that proposed system to recognize faces across various lighting conditions using Sparse Representation Technique gives the best recognition rate of 95 % on Extended Yale B database.