Reflectance hyperspectral data processing on a set of Picasso paintings: which algorithm provides what? A comparative analysis of multivariate, statistical and artificial intelligence methods

Recently a new trend towards a more systematic use of Reflectance Hyperspectral imaging (HSI) has emerged in major museums. Extensive acquisition of HSI data opens up new research topics in terms of comparative analysis, creation and population of spectral databases, linking and crossing information. However, a full exploitation of these big-size data-sets unavoidably raises new issues about data-handling and processing methods. Along with statistical and multivariate analysis, new solutions can be borrowed from the Artificial Intelligence (AI) area, using Machine Learning (ML) and Deep Learning (DL) methods. In this work different algorithms based on multivariate analysis and Artificial Intelligence methods are comparitevely applied to process HSI data acquired on three Picasso’ paintings from the Museu Picasso collection in Barcellona. By using a “data-mining approach” the HSI-data are examined to unveil new correlations and extract embedded information.

[1]  M. Picollo,et al.  Picasso’s 1917 paint materials and their influence on the condition of four paintings , 2020, SN Applied Sciences.

[2]  Susanna Bracci,et al.  The illuminated manuscript Corale 43 and its attribution to Beato Angelico: Non-invasive analysis by FORS, XRF and hyperspectral imaging techniques , 2018 .

[3]  M. Walton,et al.  Application of Uniform Manifold Approximation and Projection (UMAP) in spectral imaging of artworks. , 2021, Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy.

[4]  Matthias Alfeld,et al.  Recent developments in spectroscopic imaging techniques for historical paintings - A review , 2017 .

[5]  A. Casini,et al.  Documentation and analysis of some Picasso’s paintings by using hyperspectral imaging technique to support their conservation and stylistic matters , 2020, IOP Conference Series: Materials Science and Engineering.

[6]  David W. Messinger,et al.  Towards automatic classification of diffuse reflectance image cubes from paintings collected with hyperspectral cameras , 2020 .

[7]  A. Woollett,et al.  Rembrandt's An Old Man in Military Costume: Combining hyperspectral and MA-XRF imaging to understand how two paintings were painted on a single panel , 2019, Journal of the American Institute for Conservation.

[8]  C. Cucci,et al.  Remote-sensing hyperspectral imaging for applications in archaeological areas: Non-invasive investigations on wall paintings and on mural inscriptions in the Pompeii site , 2020 .

[9]  Matthias Alfeld,et al.  XRF and reflectance hyperspectral imaging on a 15th century illuminated manuscript: combining imaging and quantitative analysis to understand the artist’s technique , 2018, Heritage Science.

[10]  S. Kogou,et al.  From remote sensing and machine learning to the history of the Silk Road: large scale material identification on wall paintings , 2020, Scientific Reports.

[11]  Marcello Picollo,et al.  Extending hyperspectral imaging from Vis to NIR spectral regions: a novel scanner for the in-depth analysis of polychrome surfaces , 2013, Optical Metrology.

[12]  Marcello Picollo,et al.  Reflectance Hyperspectral Imaging for Investigation of Works of Art: Old Master Paintings and Illuminated Manuscripts. , 2016, Accounts of chemical research.

[13]  Paola Ricciardi,et al.  Visible and infrared imaging spectroscopy of paintings and improved reflectography , 2016, Heritage Science.

[14]  Abbie Vandivere,et al.  From ‘Vermeer Illuminated’ to ‘The Girl in the Spotlight’: approaches and methodologies for the scientific (re-)examination of Vermeer’s Girl with a Pearl Earring , 2019, Heritage Science.

[15]  Aggelos K. Katsaggelos,et al.  Innovative data reduction and visualization strategy for hyperspectral imaging datasets using t-SNE approach , 2018 .