Directional Discriminant Analysis for Image Feature Extraction

Abstract

In this paper, two novel image feature extraction algorithms base on Directional Filter Banks are proposed, which are named Single Directional Two-Dimensional Linear Discriminant Analysis (SD2DLDA) and Multiple Directional Two-Dimensional Linear Discriminant Analysis (MD2DLDA). The SD2DLDA takes least running time while achieves high recognition rate and the MD2DLDA can get the higher recognition rate than other popular algorithms. Experimental results confirm the advantages of SD2DLDA and MD2DLDA.