MARGIN-MAXIMIZATION DISCRIMINANT ANALYSIS FOR FACE RECOGNITION (MP-L1)

Author(s) : Yan Zhu (Nanyang Technological University, Singapore) Eric Sung (Nanyang Technological University, Singapore)

Abstract : LDA and its variants are popular for image-based classification problems such as face recognition. However, their performance is inherently unstable when the samples are sparse. In this paper, we propose a new type of discriminant analysis called MMDA, which derives features by maximizing the average margin between the classes. The method does not require Sw to be non-singular and well-conditioned as it does not involve its inverse term, and the features can be directly derived from the input space. A computational trick has also been proposed for MMDA to handle high-dimensional data. We conduct intensive tests on ORL and UMIST face databases, and the results show that MMDA is a good replacement of LDA for sparse sample problem.

Menu