NEW APPLICATIONS OF MULTIVARIATE STATISTICAL-ANALYSIS IN SPECTROSCOPY AND MICROSCOPY

Spectrum and image sequences are more and more often recorded in analytical spectroscopy and microscopy. They need specific processing tools. An alternative to modeling (as a function of time, energy, space coordinates) appears to be provided by multivariate statistical analysis. Possible new applications of this group of methods are explored in this paper, which concern the fields of elemental mapping by energy-filtered imaging, time-dependent spectroscopy and three-dimensional reconstruction. Besides standard uses of MSA (analysis of the data set. data compression, automatic classification, multivariate noise filtering, etc.), a new concept is introduced and emphasized: the concept of constrained synthesis, i.e., the production of "fictitious" data from a set of experimental data.