Preliminary Computational Approaches using Cluster Analysis for the Identification of Mixtures of Pigments in the case of Modern and Contemporary Paintings

Abstract

In the matter of the study of modern and contemporary paintings it is generally necessary to take samples to gain an in-depth understanding of the employed materials and techniques. However, since this procedure is characterized by its invasive nature it must be carried out in a minimal manner. This study aimed to evaluate the potentiality of color clustering applied to the corrected images of paintings to identify blends of pigments in an effort to obtain relevant preliminary information, ease the research process and guide the sampling collection. Additionally, this method would be less expensive than the traditional analytical techniques as it would only require a modified digital camera, lenses, a color target and three softwares for the processing of data, out of which two are freely available. The paintings that have been selected belong to the International Gallery of Modern Art Ca’ Pesaro in Venice (Italy) and have been depicted by three Venetian artists, namely Guido Cadorin, Andreina Rosa and Boris Brollo that date from 1921 to 1989. The artworks were thoroughly studied mainly through non-invasive analytical techniques (FORS, RAMAN, ER-FT-IR, XRF). Using clustering analysis, simulating mixtures, and calculating the color difference it was possible to infer the existence of color blends between two/three detected primary colors from the examined images, which could be validated by the analytical results. Hence, a sample taken from the color blend might suffice, as the primary colors would be concomitantly analyzed. Furthermore, it was found that paintings that consist of non-overlapping paint layers, like in the case of Boris Brollo, are the most suitable for such an analysis, as a superposition of paints might diminish the ability of the clustering algorithm to detect all colors correctly.