AI helped prove the authenticity of the controversial picture of Renoir
The Swiss company Art Recognition involved AI to verify the authenticity of the painting “Portrait of a Woman (Gabriel)” allegedly French painter Pierre Auguste Renoir. About this writes The Guardian.
Sotheby’s auction house appreciated the work belonging to a private collector at 100,000-150,000 pounds (about $ 118,000-170,000). However, art historians argue about the authenticity of the picture.
When the Vildenstein-Platner Institute, which develops Renoir’s catalogs, refused to include Gabriel in his listing, the owner of the portrait turned to Art Recognition.
The company’s specialists used photographs of 206 genuine paintings by the painter and their fragments to train AI-algorithm to recognize his smears and options for combining colors. To increase accuracy, they also trained a neural network on the works of other artists with a similar style that wrote works at the same https://gagarin.news/news/more-nf-ts-from-dubai-police-are-up-for-grabs/ time with the French impressionist.
As a result, the technology determined that with a probability of 80.58%, the “portrait of a woman” belongs to the brush of Renoir.
After that, the owner of the picture turned to the Parisian experts GP.F.Dauberville&Archives Bernheim-Jeune, publishing its own catalogs of Renoir’s works. After conducting a scientific analysis of the pigments of the portrait, they also confirmed the authenticity of Gabriel.
However, connoisseurs of art are sure that the accuracy of AI depends on the quality of the paintings on which he studied. If the works are fake or contain tinted areas, this will affect the operation of the system and create even greater uncertainty.
Art critic Bendor Grosvenor expressed concern that such technologies can depreciate the contribution of experts to assess the authenticity of the paintings.
“So far, the methods used to teach II programs, and the fact that the company judges attributes only by photography with the iPhone is not impressive,” he said.
Grosvenor added that the technology is weak in its inability to take into account the state of works.
“Many paintings of old masters are damaged and mutilated with layers of dirt and paint, and without a forensic examination it is impossible to determine whether the work is the original or not,” he added.
According to the head of Art Recognition, Karina Popovichi, the company understand the importance of the quality of training datasets, so they used photos of only genuine paintings.
“We sincerely want to give art critics the opportunity to apply AI in order to help make decisions in cases where they doubt. But they must be open to this technology, ”she said.
Recall that in October 2021, the ART Recognition developed by the authenticity of the Rubens Rubens “Samson and Dalila”, represented in the National Gallery in London.
In the same month, Google Ai specialists recreated the destroyed paintings of Gustav Klimt using artificial intelligence.
Subscribe to FORKLOG news at Telegram: Forklog AI – all news from the world of AI!