As visitors flock back to Italy’s museums and galleries, cutting-edge AI technology is helping one institution gauge the popularity of its artworks. The data compiled through the project hopes to aid this museum and others with the curation and organization of their collections. 

AI Watches Visitors Watch Art

At Istituzione Bologna Musei, an art gallery in the northern Italian city of Bologna, a little device sits above a painting display label. Its camera is recording the number of visitors who observe the painting and their behavior, without compromising privacy. It is part of a trial of new technology developed by the agency ENEA. With it, they hope to assist museums in assessing the attractiveness of exhibits in their collection. 

The ShareArt devices attached close to the museum’s artifacts calculate both how long and how closely visitors observe an exhibit. As a result, the museum can begin to understand which paintings, sculptures or artifacts are the biggest draws. 

The ShareArt system was conceived back in 2016 by ENEA researchers. However, trials in museums only began a few weeks ago when Italy’s government gave the green light for the reopening of cultural institutions closed amid the coronavirus emergency.  

The resulting data can help subsequently the museum reorganize the collection, change gallery layouts or alter the exhibits on display according to what interests visitors. 

What Makes Visitors Smile? 

At the presentation of the project, the ENEA team explained, “Through a camera positioned near the work, the system automatically detects faces looking in the direction of the work itself, acquiring data relating to the behavior of observers such as, for example, the path taken to approach, the number of people who have observed, the time and distance of observation, the gender, the age class and the state of mind of the visitors who observe.”

As a result, the devices monitor where an observer’s eye focuses on an artwork and their emotional reaction. Currently, ShareArt can distinguish five expressions: happy, sad, neutral, angry or surprised. 

At Istituzione Bologna Musei, the company has installed 14 ShareArt devices. They have been allocated to a variety of artifacts of different natures and dating from different periods. The ShareArt camera translates the way visitors are interacting with these exhibits into a graphic. As such, the ENEA team and the museum can see what draws visitors’ attention. 

Collections Curated by AI

“There are questions that run around within the walls of a museum. What does the appreciation of an artwork consist of? What are the personal and environmental variables that affect this appreciation? The traditional answers are too approximate”, said Roberto Grandi, president of Istituzione Bologna Musei, at the project’s presentation.

The findings of ShareArt have already proved interesting. The data has shown that on certain paintings visitors tend to focus on a particular part of a composition. In one case, they often skipped one side of a diptych. In general, ShareArt has found that visitors rarely spend more than four or five seconds before an artwork. 

These kinds of findings may reveal to the museum that the layout of exhibits is causing visitors to act in a certain way. Perhaps observers tend not to look at the other side of the diptych because something catches their eye and draws them away. Galleries may also learn more about how their lighting or display cases influence visitor behavior. 

In due course, and when mask mandates lift, ENEA is considering advanced techniques to observe more facial expressions, too. The researchers hope to analyze visitors’ reactions with cognitive psychology to further determine the effect of a collection on observers.