A Mixed Two-dimensional Linear Discriminate Method

Abstract

Feature extraction is one of key technologies of the palmprint identification. In the light of the characteristics subspace palmprint identification technology, the two-dimensional principal component analysis, two-dimensional fisher linear discriminant and two-way two-dimensional principal component analysis algorithm is deeply analyzed. Based on two-dimensional subspace palmprint identification algorithm is a direct projection of the palmprint image matrix and is achieved very good results for dimension reduction. This paper proposed a mixed two-dimensional linear discriminant dimension reduction algorithm which can eliminate the relevance of rows and columns to get the best projection vector and extract optimal discriminant information. Experimental results show that the proposed method has faster extraction speed, higher recognition rate and better robustness. DOI: http://dx.doi.org/10.11591/telkomnika.v11i6.2608