Alex Kachkin, a graduate student at the Massachusetts Institute of Technology (MIT), USA developed a tool based on artificial intelligence, capable of quickly restoring damaged paintings.
“Restoring a painting is fun and it’s great to sit down and fill it all in and have a nice evening. But it’s a very slow process.” – said the specialist.
According to him, this method can radically change the approach in this area. He noted that in the past such restoration work took months or even years of painstaking work.
The new technology is based on the creation of a digital reconstruction of the work, where damaged areas are restored by AI based on the analysis of neighboring fragments. After that, the resulting mask is printed on a polymer film and carefully applied to the original. If necessary, the film can be removed without harming the canvas itself, which makes the process completely reversible.
Kachkin tested his technique on a 15th-century painting attributed to the so-called Prado Master of the Adoration of the Magi (the true identity of the artist remains unknown). He digitized the images, marked more than 5,600 defects and restored them using a software mask containing more than 57,000 unique colors. Fragments from other works of this author were used to restore the lost elements, in particular the baby’s face. This is how the AI picked up the necessary details after studying the artist’s creative style.
The entire process took only a few hours instead of the expected eight days for a traditional restoration. Although the technology is suitable only for varnished canvases, Kachkin believes that it will allow many forgotten masterpieces to be returned to exhibitions.
However, along with new opportunities, ethical issues also arise, the author of the tool emphasized. Is it permissible to cover the original with a film and is it ethical to reproduce the lost fragments, especially if they are borrowed from other works?
“It will take a lot of thinking about the ethical issues that arise at each stage of this process to understand how this can be applied in a way that is as consistent as possible with conservation principles. We are laying the groundwork for the development of further methods. As others work on this, we will arrive at more precise methods.”, – noted Kachkin.
