Utilizing AI to elucidate the chemical composition of paints used in classical paintings
Restoration of the Baronchi Altarpiece by Raphael. In this work, the two paintings at the top, the Virgin Mary (left) and God the Father (right), are analyzed. The two paintings belong to the collection of the Capodimonte Museum (Naples). On the lower left is an angel holding a holy verse (Paris, Louvre Museum), and on the lower right is a statue of an angel (Tosio Martinengo Gallery, Brescia). Credit: Danilo Pavone, ISPC-CNR, Catania
A team of chemists and AI researchers at the CNR Institute of Science and Culture has developed an AI model that can determine the chemical composition of the paints used to create classical paintings.
In a paper published in Science Advances, the group describes how they developed and trained an AI model using a dataset containing information on 500,000 synthetic spectra representing 57 dyes and related compounds. I’m explaining.
Maintaining and restoring old paintings, especially valuable paintings, is both an art and a science. Our experts are trained in a variety of fields, from chemistry to botany to history. Due to the high value of such works of art, new techniques are being explored to better understand the nature of a particular painting before any restoration work takes place.
One of the main areas of interest is the chemical composition of the paints used by artists. Using the wrong chemicals can cause a reaction that destroys the paint and ruins your ancient masterpiece. In this new effort, the research team brought artificial intelligence to the challenge.
High-resolution MA-XRF elemental distribution map showing details of God the Father. (A) Visible image of the scanned area. (B) Elemental map of Pb-L. (C) Elemental map of Hg-L. (D) Red-green-blue (RGB) composite image of elemental distribution maps of Hg-L, Fe-K, and Cu-K. This image provides insight into Raphael’s painting techniques. (E) Elemental distribution map of Pb-M. AI/ML estimates are lower than reference values. However, the network provides a less noisy output. (F) Elemental distribution map of SK. The AI/ML image is predicted, but the reference image is too noisy to interpret due to the low net number of sulfur and the energy overlap of SK and Pb-M. Credit: Photo (A) by Danilo Pavone, ISPC-CNR, Catania; Scientific Progress (2024). DOI: 10.1126/sciadv.adp6234
To understand the chemicals that make up a particular paint, experts use fluorescent X-rays. X-ray imaging is a non-invasive method that provides detailed elemental composition related to the paint used in a particular painting. . Unfortunately, the fact that artists mix pigments to obtain the desired color makes it even more difficult to identify individual paints.
Attempting to identify the chemicals in such mixtures often requires educated guesses and errors are made. To reduce such errors, researchers received and analyzed macro X-ray fluorescence (MA-XRF) datasets to determine the chemistry present in all the oils used to create a particular painting. We have developed an AI model that can print out materials. The model was trained using a dataset containing information on 500,000 synthetic spectra.
Once the model was complete and initial tests were completed, the researchers conducted more realistic tests to determine which chemicals were present in the oil used to create two paintings by the artist Raphael, painted between 1501 and 1502. asked to specify.
Both have been extensively researched and tested using other methods, so their chemical components are pre-identified. The researchers found that the model was able to accurately identify chemicals such as lead in white paint, mercury in red paint, and copper in green paint.
More information: Zdenek Preisler et al., Deep learning for enhanced spectral analysis of painting MA-XRF datasets, Science Advances (2024). DOI: 10.1126/sciadv.adp6234
© 2024 Science X Network
Citation: Using AI to understand the chemical composition of paints used in classical paintings (October 5, 2024) https://phys.org/news/2024-10-ai-figure-chemical-composition- Retrieved October 5, 2024 from classical.html
This document is subject to copyright. No part may be reproduced without written permission, except in fair dealing for personal study or research purposes. Content is provided for informational purposes only.
