Feature extraction based on two-directional diagnal scatter difference analysis

Abstract

In this paper we propose a novel method for image feature extraction. The novel method is developed based on two techniques, i.e., maximum scatter difference (MSD) and diagonal face images based projection. In addition, an automatic strategy for selecting the 2D components or discriminative features is adopted. Extensive experiments are performed to test and evaluate the new algorithm using the AR face databases. Experimental results show the effectiveness of the proposed method.