Computer curators: Artificial intelligence learns art history

Crossovers between art and the sciences are making a huge impact on the art landscape.

In September 2012, computer scientists Lior Shamir and Jane Tarakhovsky of Detroit’s Lawrence Technological University published the results of an experiment, a computer programme that can classify paintings according to their art historical period or movement.

The computer-generated groupings of painters from art history. Image courtesy Lior Shamir.

The programme uses 4,027 image content descriptors to quantify the visual features of each painting, including things like texture, colour and shape. With these metrics, the computer algorithm was able to correctly group approximately 1,000 artworks from 34 artists based on their visual styles, with no mistakes.

According to the study, the programme immediately divided the 34 artists into two groups, classical and modern painters. From there, the algorithm was also able to identify sub-groups whose artists belonged to the same art movements, linking Da Vinci with Michaelangelo, Rubens with Vermeer and Gauguin with Cézanne.

Though the programme represents a major milestone for computerised visual analysis, it will probably end up being more useful to an untrained viewer of art than to art historians.

While the average non-expert can normally make the broad differentiation between modern art and classical realism, they have difficulty telling the difference between closely related schools of art such as Early and High Renaissance or Mannerism and Romanticism. The experiment showed that machines can outperform untrained humans in the analysis of fine art.

Will computers take the place of art historians in the not-too-distant future? Leave your thoughts in the comment section below.


Related Topics: research, art and the science collaboration, art and technology

Related Posts:

Subscribe to Art Radar for more about art and technology

Leave a Reply

Your email address will not be published.