Various multivariant analysis (MVA) methods have been used to denoise [1] or extract physically interpretable information [2-4]. The idea is to project the dataset on a subspace of lower dimension,… Click to show full abstract
Various multivariant analysis (MVA) methods have been used to denoise [1] or extract physically interpretable information [2-4]. The idea is to project the dataset on a subspace of lower dimension, which, in the ideal case, is described by physically meaningful spectra. MVA methods assume that any spectrum is a linear combination of a given set of spectra but does not account for the redundant information in neighbouring pixels. Indeed, MVA techniques look for spectral rather than spatial similarities.
               
Click one of the above tabs to view related content.