Scientists have developed an AI-based algorithm for analyzing paintings that allows determining whether a canvas belongs to a certain master

Scientists have taught artificial intelligence (AI) to see image details that are hidden from the human eye. Thanks to this, the neural network was able to establish that one of the most famous paintings by Raphael is not one hundred percent his work, writes UNN with reference to a study published in Heritage Science.

We are talking about the painting "Madonna with a Rose" (another title "The Holy Family with St. Joseph"), probably painted in 1518-1520. This canvas has long been a source of controversy among art historians – some experts doubted whether this painting was really created by Raphael.

Thanks to the use of AI, the researchers were able to find out that one of the faces in the painting, namely the face of Joseph, was painted by another artist. But the Madonna herself, baby Jesus, and St. John were painted by Raphael.

Researchers from the United Kingdom and the United States have developed a special analysis algorithm based on works that we know to be the result of the work of the Italian master.

"Using deep feature analysis, we used photographs of authenticated paintings by Raphael to train the computer to recognize his style in great detail, from brushstrokes, color palette, shading, and every aspect of the work," said mathematician and computer scientist Hassan Ugail of the University of Bradford in the United Kingdom.

In their work, the researchers used a modified neural network developed by Microscoft called ResNet50, combined with a traditional machine learning technique called Support Vector Machine. This method demonstrated 98 percent accuracy in identifying Raphael's work.

In the future, the researchers hope to use their findings to help art historians identify works of art.

For decades, Picasso's painting hung in the living room of a dealer who had no idea about the artistOctober 1 2024, 06:01 PM • 104045 views