AI Restores Old Art: New Tool & Science Breakthrough
- Centuries of wear and tear can leave oil paintings cracked, discolored, and missing pigment.
- The process involves using AI and computer tools to create a digital reconstruction of the damaged painting.
- Alex Kachkine, a graduate researcher at MIT, demonstrated the technique by restoring a damaged oil-on-panel work attributed to the Master of the Prado adoration, a 15th-century Dutch painter.
Artificial intelligence is revolutionizing art restoration, and this new technique offers a much faster, cheaper way to repair damaged paintings. AI can now restore aging masterpieces in just hours, using a digital reconstruction printed on a clear polymer sheet. This innovative approach to art restoration allows for more precise repairs, expanding the potential for museums and galleries to showcase works that would otherwise remain hidden. News Directory 3 reports on the technique thatS already breathing new life into centuries-old oil paintings once deemed beyond repair. Discover what’s next for this groundbreaking technology.
AI-Powered Art Restoration Offers Hope for Damaged Masterpieces
Updated June 11, 2025
Centuries of wear and tear can leave oil paintings cracked, discolored, and missing pigment. While customary restoration can take years and is reserved for the most valuable works,a new approach using artificial intelligence promises to restore aged artworks in mere hours. This innovative technique offers a faster and more cost-effective solution for art conservation.
The process involves using AI and computer tools to create a digital reconstruction of the damaged painting. This reconstruction is then printed onto a transparent polymer sheet, which is carefully applied to the original artwork. This method of art restoration allows for precise and detailed repairs,bringing new life to aging masterpieces.
Alex Kachkine, a graduate researcher at MIT, demonstrated the technique by restoring a damaged oil-on-panel work attributed to the Master of the Prado adoration, a 15th-century Dutch painter. The painting, visibly split into four panels and covered in cracks and missing paint, served as an ideal test case for the AI restoration method.
Kachkine estimated that traditional conservation techniques would have required about 200 hours to restore the painting. “A lot of the damage is to small, intricate features,” Kachkine said, noting the centuries of degradation the artwork had undergone.
The process began with a scan of the painting to identify the size, shape, and position of the damaged areas, revealing 5,612 separate sections needing repair. A digital mask was then constructed using Adobe Photoshop. Missing specks of paint were restored,and colors were matched to surrounding pigments. Damaged patterns were corrected by copying similar patterns from elsewhere in the painting. In one instance, the missing face of an infant was copied from another work by the same artist. In total, 57,314 colors were used to infill damaged areas, improving the painting even if the corrections were not perfectly aligned. The digital art restoration mask was then printed on a polymer sheet and varnished before being overlaid on the painting.
“It followed years of effort to try to get the method working,” said Kachkine. “there was a fair bit of relief that finally this method was able to reconstruct and stitch together the surviving parts of the painting.”
The technique, detailed in Nature, is suitable for varnished paintings with smooth surfaces. The mask can be safely removed using conservators’ solvents, leaving no trace on the original artwork.
What’s next
Kachkine hopes this method will enable galleries to restore and display numerous damaged paintings that are not currently deemed valuable enough for traditional restoration. While ethical considerations remain, such as the acceptability of a film covering a painting and the appropriateness of certain corrections, the potential to widen public access to art is important.