Dimensionality Reduction, Classification, and SpectralMixture Analysis using Nonnegative Underapproximation

摘要

Nonnegative Matrix Factorization (NMF) and its variants have recently been successfully used as dimensionalityreduction techniques for identification of the materials present in hyperspectral images. In this paper, we presenta new variant of NMF called Nonnegative Matrix Underapproximation (NMU): it is based on the introductionof underapproximation constraints which enables one to extract features in a recursive way, like PCA, butpreserving nonnegativity. Moreover, we explain why these additional constraints make NMU particularly well-suited to achieve a parts-based and sparse representation of the data, enabling it to recover the constitutiveelements in hyperspectral data. We experimentally show the efficiency of this new strategy on hyperspectralimages associated with space object material identification, and on HYDICE and related remote sensing images.