Skip to main content


Innovative Data Analyses Tool for Hyperspectral DataCube


Visible hyperspectral imaging (HSI) is a fast and non-invasive imaging method that has been adopted by the field of conservation science to study painted surfaces. By collecting reflectance spectra from a 2D surface, the resulting 3D hyperspectral data cube contains millions of recorded spectra. While processing such large amounts of spectra poses an analytical and computational challenge, it also opens up new opportunities to apply powerful methods of multivariate analysis for data evaluation.

Innovative Solutions for Non Linear Unmixing

With the intent of expanding current data treatment of hyperspectral datasets, and solving the problem of nonlinear unmixing of hyperspectral reflectance data acquired on painted works of art, two innovative data analysis approaches have been recently developed. First, a data driven approach for data reduction and visualization that uses a statistical embedding method known as t-distributed stochastic neighbor embedding (t-SNE) successfully provides a non-linear representation of spectral features in a lower 2D space. Second, a nonlinear unmixing approach using overdetermined pigment dictionary combined with supervised classification and two-constant Kubelka-Munk theory efficiently determines pigment concentrations within the pixel. The efficiency and limitations of the proposed methods for painted surfaces from cultural heritage are currently evaluated through the study of laboratory prepared paint mock-ups, and medieval French illuminated manuscript, part of the Isabella Gardner Museum collection (Boston, USA).


View All Projects

Back to top