Neighborhood Discriminant Nearest Feature Line Analysis and Its Application to Face Recognition

Abstract

In this paper, a novel subspace learning algorithm, called neighborhood discriminant nearest feature line analysis (NDNFLA), is proposed. NDNFLA aims to find the discriminant feature of samples by maximizing the between-class feature line (FL) scatter and minimizing the within-class FL scatter. At the same time, the neighborhood is preserved in the feature space. Experimental results demonstrate the efficiency of the proposed algorithm.