Neural network simulation of original colors in Friedrich’s Abbey Among Oak Trees featuring discoloured smalt

Artwork appearances change over time due to aging. Smalt, a blue cobalt-tinted glass pigment, deteriorates over time in oil paintings causing significant and irreversible color changes in many artworks. Virtual simulations can hypothesis original appearances while it remains a challenge for smalt-co...

Full description

Saved in:
Bibliographic Details
Published in:Heritage science Vol. 13; no. 1; p. 388
Main Authors: de Mecquenem, Clément, Eveno, Myriam, Alfeld, Matthias, Calligaro, Thomas, Laval, Eric, Mösl, Kristina, Reiche, Ina
Format: Journal Article
Language:English
Published: London Springer Nature B.V 06.08.2025
Subjects:
Aging
Art works
Artificial intelligence
Cobalt
Color
Discoloration
Electron microscopes
Friedrich, Caspar David (1774-1840)
Machine learning
Neural networks
Oil painting
Paints
Pigments
Scanning electron microscopy
Speciation
Spectrum analysis
Tinted glass
X ray absorption
X-ray fluorescence
ISSN:2050-7445
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Artwork appearances change over time due to aging. Smalt, a blue cobalt-tinted glass pigment, deteriorates over time in oil paintings causing significant and irreversible color changes in many artworks. Virtual simulations can hypothesis original appearances while it remains a challenge for smalt-containing paintings. A novel procedure integrates non-invasive imaging methods, X-ray absorption near-edge structure (XANES), and machine learning to simulate the original colors of a smalt-containing discolored paintings. Macro-X-ray fluorescence provided elemental distribution, reflectance imaging spectroscopy captured color spectra of pigments and XANES informed cobalt speciation in cross sections. Friedrich’s Abbey Among Oak Trees (1808-1810) containing smalt and artificially aged model systems were studied. Machine learning predicted the original hues based on XANES. The procedure allowed us to simulate the original, cooler and more vibrant colors of the painting. The innovative approach visualizes a possible original state of the smalt-containing artwork that can be adapted to other alteration phenomena.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2050-7445
DOI:10.1038/s40494-025-01953-y