New Tech Can Distinguish Brush Strokes of Different Artists

artist's palette with brush
The new technique can distinguish artists based on small samples of their brushwork. Pxhere under CC0 Public Domain

A new artificial intelligence (A.I.) tool may be able to foil fraud and help art historians determine the original creator behind particular paintings. The system analyzes tiny sections of paintings, some as small as half a millimeter, for telltale differences in brushwork, reports Benjamin Sutton for the Art Newspaper.

While previous projects used a form of machine learning to identify artists based on the analysis of high-resolution images of the paintings, the new system uses topographical scans of the canvasses.

“We found that even at the brush bristle level, there was a fair level of success in sorting the attribution,” Kenneth Singer, a physicist at Case Western Reserve University, tells the Art Newspaper. “Frankly we don’t really understand that, it’s kind of mind boggling actually when you think about it, how the paint coming off a single bristle is indicative of what we’re calling the artist’s unintentional style.”

The research is a result of a collaboration several years ago between Michael McMaster, then a physics graduate student working with Singer, and Lauryn Smith, an art history scholar. With Singer and other colleagues, the pair published their findings last November in the journal Heritage Science.

To test the A.I. system, four art students at the Cleveland Institute of Art each painted yellow flowers using identical brushes, paints and canvases, reports Steven Litt reports for Cleveland.com. The researchers scanned the surfaces of the paintings using a tool known as a chromatic confocal optical profilometer, creating precise 3-D surface height data showing how the paint lay on the canvases, and digitally broke them into grids. The machine-learning system analyzed randomized samples and was able to sort them by the artist with a high level of accuracy. 

“We broke the painting down into virtual patches ranging from one-half millimeter to a few centimeters square, so we no longer even have information about the subject matter,” says Michael Hinczewski, another Case Western physicist and coauthor of the study, in a statement. “But we can accurately predict who painted it from an individual patch. That’s amazing.”

Grid showing painted images and 3D scans
The researchers digitally broke the 3-D scans of the images into small sections. Heritage Science

In additional research not yet published, the team used the A.I. to try to distinguish original portions of the 17th-century painting Portrait of Juan Pardo de Tavera (1609) by El Greco from sections that were damaged during the Spanish Civil War and restored later. 

“This is a painting we have an answer key to, because we have photos of the destroyed painting and the current painting, so we’re able to make a map of the areas that were conserved, and [the A.I.] was able to identify those areas,” Singer tells the Art Newspaper. “But there was another section of the painting that it identified as conserved that wasn’t obvious, so we’re going to have a painting conservator in Spain look at the painting to see what’s going on.”

The team’s next project is analyzing two paintings of the crucifixion of Christ by El Greco in the hopes of distinguishing portions painted by himself, by his son Jorge Manuel; by other members of his workshop; and by later conservators.

“The El Greco project is looking at several different scans of paintings to see if we can identify the workshop process and identify different hands,’’ Elizabeth Bolman, an art historian and coauthor of the paper, tells Cleveland.com. “Did he work on them? How much did his son Jorge work on them? These are hotly contested issues.’’

The workshop system employed by El Greco was used by many of the European Old Masters, according to the John and Mable Ringling Museum of Art. Starting in the 15th century, master artists began bringing together students and assistants to produce work for the market more quickly, imitating the star artist’s style. The system was especially popular in the 17th century, when artists including Peter Paul Rubens and Rembrandt had large teams helping them in their work.

The new technology may help art historians tease out details of how these collaborations worked, as well as helping to determine the authenticity of work for sale on art markets.

“We’re at the point where we’ve just figured out the basics of a concept and our first attempt ended up being spectacularly successful beyond our wildest dreams,’’ says Bolman. “Where this goes from here, we can all dream.”

Comments are closed.