No abstract

No abstract

No abstract

No abstract

Linear subspace learning has achieved great success in feature extraction, and it aims to map high dimensional data into low dimensional feature space which can reflect the important inherent structure of original data. In this paper, a novel approach termed Discrimination Preserving Projection (DPP) based on Sparse coding is proposed, which mainly focus on combining locality supervised linear subspace learning with sparse coding. In our approach, we decompose images into two parts including more discrimination part and less discrimination part via dictionary learning and sparse coding firstly. Then, a locality supervised criterion which preserves the more discrimination part components while weaken the less discrimination part components is presented. Extensive experiments on publicly available databases are conducted to verify the effectiveness of the proposed algorithm and corroborate the above claims.