Discover how a team of scientists is leveraging the power of artificial intelligence to unravel the chemical composition of paints used in classical paintings. X-ray fluorescence and painting restoration techniques are combined to preserve the integrity of these masterpieces.
Chemical Secrets Of Classical Paintings Unlocked By AI
Engineers and Data Scientists at CNR-Istituto di Scienze del Patrimonio Culturale have made a phenomenal breakthrough in developing an AI model that enables the scientists/chemists to find out which paints are used with classical paintings. This new method is poised to redefine painting restoration and conservation.
The restoration of paintings is a discipline that can be compared to an art and a science and which requires knowledge of chemistry, botany, and art history. The key to this is understanding the chemical nature of the paints used by the artist. Careless application of chemicals in the restoration process can cause reactions that could destroy the master piece. To combat this difficulty, the team of researchers has tapped into artificial intelligence for a far more accurate and effective response.
Using X-ray Fluorescence, AI to Reveal Paint Mixtures
To that end, the scientists used a non-invasive analysis method called macro X-ray fluorescence (MA-XRF) to learn about the composition of the paints. The technique has the advantage of being able to examine the chemical nature of the paint used in detail and is non-destructive to the artwork itself.
This has historically been the case with some paint mixtures of artists, where identification of a single pigment is complex. the team designed an AI to break down the MA-XRF datasets and then determine what chemicals were in the oil used for who painting This data set was used to train a model on more than 500,000 synthetic spectra (both live and crushed) which encompassed examples of all the successful pigments and dye compounds described in this study — 57 precursors in total.
The research team tested the AI model by trying to determine chemical compounds in oils that were used in paints when he creating 2 paintings, painted with oil colors of the famous painter Raphael (from 1501-1502). When fine tuning the AI on these paintings, extensive existing work and studies (using other methods) provided a highly reliable baseline against which to measure the performance of the AI.
AI Chemical Analysis Innovating Painting Restoration
According to a company blog posts about the study, the results were surprisingly life-changing. The AI model could also identify the chemicals found in the Raphael paintings, to be lead from white paint, mercury from red paint and copper from green paint. This kind of precision represents a major advancement for painting restoration and conservation in general.
An example of use for experts in the field who are responsible for preserving, protecting and possibly restoring classical paintings, that is what this research team has accomplished using AI. This technology could provide important information to guide that policy, and help them learn how best to treat the ancient chemicals of pigments for priceless works of art.
It is inspiring to see how this AI technology can be used in the preservation of our culture as AI exponentially grows. However, the researchers also show that their methods have general applicability beyond art history and could be used in other domains of culture, as well as in studies of history. We really are experiencing the first days of a new era, in terms of preserving and engaging with our artistic heritage.
