Developing a Framework for the in-situ Analysis of Artists’ Prints and Paintings via Hyperspectral Imaging

Hyperspectral imaging is a novel technology which allows for the collection of reflectance spectra of a sample in-situ and at a distance. A rapidly developing technology, hyperspectral imaging has been of particular interest in the field of art characterization, authentication, and conservation as it avoids the pitfalls of traditional characterization techniques and allows for the rapid and wide collection of data never before possible. It is hypothesized that by combining the power of hyperspectral imaging with machine learning, a new framework for the in-situ and automated characterization and authentication of artworks can be developed. This project, using the CMYK set of inks, began the preliminary development of such a framework. It was found that hyperspectral imaging and machine learning as a combination show significant potential as an avenue for art authentication, though further progress and research is needed to match the reliability of status quo techniques.