WILL COMPUTERS BECOME PERFECT ART FORGERS?

The choice of Rembrandt was no coincidence. The reason: “There are many images by the artist,” says art historian and materials scientist Joris Dik of Delft University of Technology in the Netherlands. In fact, computers can only learn the style of highly productive artists— and their output is not (yet) 100% perfect: Certain details, such as the light reflected onto the chin by the collar, are missing. “But the technique could soon help expose forgeries when it comes to newly discovered paintings,” says art historian Gary Schwartz. When utilized in such an instance the software would compare all “real” noses painted by Rembrandt with the structure of the new one and evaluate their similarity.