Tensor decomposition for painting analysis. Part 2: spatio-temporal simulation

Abstract In a previous article, we modelled the spectral and temporal dimensions of photodegradation behaviour pigments in painting “A Japanese Lantern” by Oda Krohg. particular, extracted endmembers fading rate applying tensor decomposition on time-series spectroscopic point measurements. Now, capture same with hyperspectral imaging setup propose an approach to render effects as 2D images. More precisely, from image, compute concentration maps each previously identified endmember least-squares unmixing method. Subsequently, using algebra, multiply their corresponding obtain 4D where pixel image is described spectrum function. This way, generate past future spatio-temporal simulations painting’s appearance reversing elevating light exposure, respectively.