Joint data treatment for Vis–NIR reflectance imaging spectroscopy and XRF imaging acquired in the Theban Necropolis in Egypt by data fusion and t-SNE

Abstract

Hyperspectral imaging allows the acquisition of representative information on cultural heritage objects. For a complete identification of the chemical compounds present, one needs instruments employing different forms of radiation, sensitive to different chemical features. Data treatment workflows are commonly developed for a single spectroscopic technique, so that data sets are evaluated separately, and the reduced representations yielded are compared during interpretation. In this publication, we describe how t-stochastic neighbor embedding (t-SNE) allows for the evaluation and interpretation of data acquired in a Ramesside tomb (13th century BC) of the Theban Necropolis in Egypt. By fusing X-ray Fluorescence (XRF) imaging and Reflectance Imaging Spectroscopy (RIS) into a single data set, we exploit the synergies between both methods to support interpretation and presentation of the results yielded.