MIT engineering students use AI to restore famous paintings

MIT engineering student Alex Kachkine has developed a groundbreaking AI-driven method for artwork restoration, drastically reducing the traditional restoration process to just a few hours. His research shows that the more damage a painting has sustained, the more effective the new method becomes. The process involves cleaning and scanning the original artwork, conducting analytical assessments, creating dual-layer masks, and applying them to the painting’s surface using high-fidelity printing technology. In one case, Kachkine successfully restored a 15th-century oil painting with 5,612 damaged areas in just 3.5 hours—66 times faster than conventional methods. This interdisciplinary achievement not only enhances restoration efficiency but also opens up new possibilities for recovering more damaged artworks.