High-dimensional discriminant analysis and its application to color face images
Zhizheng Liang · 2006
Multilinear algebra plays an important role in signal processing. In this paper, a multilinear generalization of linear discriminant analysis is discussed. Similar to HOSVD, an iterative algorithm is developed for solving high-dimensional linear discriminant analysis. Meanwhile, a non-iterative algorithm is also proposed. In addition, HDLDA provides a unified framework for classical LDA and 2DLDA. Experimental results on color face images show the effectiveness and usefulness of the proposed method.