The use of artificial intelligence in visual art and art, in general, is going even further. After an algorithm that turns crappy sketches into works of art, through functional AI doing their own art exhibitions, to algorithms turning classical works of art into full-fledged videos, it has now come a time to use AI in art history and discovering connections between artists and the works of art they themselves used as inspiration.
As MIT Technology Review reports, two researchers at the Czech Technical University in Prague have used a machine vision system that can analyze poses of human subjects in fine art paintings throughout history. The system would then make a search for other artworks with people in the same poses.
MITTR adds that at Carnegie Mellon University in Pittsburgh, the researchers have developed algorithms that can determine a human pose from a 2D image. Also, researchers have already used algorithms so that they could find new links between artworks.
As for their research, Janicek and Chum’s goal is to search for similar poses in a database of manually annotated images. “They go on to look for similar poses in an online database called the Web Gallery of Art, which contains 37,000 images. The researchers say their algorithm discovered a wide range of links between pictures that would have been impossible to identify by other means.”
As they put it, “We experimentally show that explicit human pose matching is superior to standard content-based image retrieval methods on a manually annotated art composition transfer dataset.”