AI art: has science unravelled how we see and appreciate art?

Art, it’s in the eye of the beholder but has science found a way to identify what we really appreciate in paintings’ The artistic tastes of individuals may have been finally unravelled thanks to University of Bristol researchers, artificial intelligence (AI) and eye-tracking technology.

Delving into the mindset of an individual’s artistic preferences has proven challenging as a systematic approach to assess individual preferences towards different artistic styles and the semantic contents of paintings in an objective way has been lacking. For example, does an individual observer find paintings of Van Gogh pleasing because of the application of brushstrokes or the choice of content such as sunflowers?

A team from Bristol’s School of Psychological Sciences has conducted an experiment using artificial intelligence to create a series of images with the same content but in the artistic style of Van Gogh or CÚzanne. Volunteers were fitted with eye tracking devices pinpointing where their gaze was held while they were shown paintings either in the same style but different content, or the same content and different style. Participants were then asked to choose the painting they preferred.

The eye-tracking data, which reflected the strength of the participant’s preference for certain styles or contents in a painting, allowed researchers to analyse and identify what elements made the painting more visually appealing to the individual.

Dr Jasmina Stevanov, project lead and Senior Research Associate in Bristol’s School of Psychological Science, said: “Individual differences in art appreciation are vast but we have found a way that can predict future choices for paintings with similar style or portraying the same subject. We are working on a software algorithm that will provide automatic feedback about observers’ aesthetic preferences by looking at how people look at paintings. This is a crucial step toward personalised exhibition tailoring.”

More volunteers are now needed to take part in the researchers’ short eye-tracking experiment at their ‘Painting and Curating with AI’ exhibition. The exhibition features Jasmina Stevanov’s art photographs turned into artificial paintings with style transferred from various painters. The free event takes place at The Gallery Space, The Island, between Tuesday 3 to Sunday 8 September 2019 between 12 noon and 8 pm.

In partnership with VR-All-Art , The Island Gallery has also been replicated in Virtual Reality, giving Bristol’s School of Psychological Sciences yet another innovative method, a new tool to conduct their research. VR-All-Art is a platform and a marketplace for galleries, artists, collectors, and museums to showcase, buy and sell art primarily in Virtual and Augmented Reality. The company’s mission is to create a more immersive experience for art buyers and give new advantages to art sellers, changing the way we present, discover, experience, and trade art. Studying how art affects people in Virtual Reality is the next step in the reaserch as the technology now provides us abilities to collect large data sets from experiments in replicated Bristol’s Island Gallery Space in VR.

The research is funded by Bristol Vision Institute’s EPSRC Platform Grant - Vision for the Future. For further information about this and the event, please contact Dr Jasmina Stevanov, email, jasmina.stevanov [at] bristol.ac (p) uk.