Artificial intelligence helped a team of researchers analyze about a million images of living rooms all over the world to better understand how different cultures decorate their homes.
According to a Penn State news release, the researchers used a type of machine learning, called transfer learning, to pick out wall colors, wall art, books, and other beautifying techniques to see how people personalized their living rooms — about 50,000 living rooms, as a matter of fact. The living room pictures were collected from the Airbnb website, a place where people market their rental homes and rooms.
“We were interested in seeing how other cultures decorated,” Clio Andris, assistant professor of geography, Penn State and an Institute for CyberScience associate, said in a Penn State ICS news release. “We see maps of the world and wonder, ‘What’s it like living there,’ but we don’t really know what it’s like to be in people’s living rooms and in their houses. This was like people around the world inviting us into their homes.”
In some cases, the researchers found clusters of design elements in certain areas that matched expectations for how cultures might decorate in those regions, according to Xi Liu, doctoral student in geography, Penn State and lead author of the study, which was published in EPJ Data Science.
“There were a lot of vibrant colors in India and Morocco, for example,” said Liu. “And, of course, that wasn’t a big surprise — we had an idea that this might be the case before we started the study, but we weren’t sure whether that would be true or not.”
On the other hand, some places didn’t mesh with the styles most people associate with them. For instance, when we think of the Caribbean islands, we tend to think of vibrant colors and a lot of knickknacks and other fun decorations. But, the researchers found the islands were pretty modest compared to the lush interiors portrayed in all the travel brochures. It also turns out that even though plants may be at a premium in Scandinavian countries, the people there were willing to pay a premium to use indoor plants to warm up their interior spaces.
By using economic data from the U.S. Census to examine six U.S. cities on a neighborhood level, the researchers also saw that the desire to decorate runs across socioeconomic barriers. People in neighborhoods with varying incomes, unemployment rates, educational attainment, residential property value and racial diversity all showed about the same level of decorative efforts in their homes.
Because picking out wall paintings and house plants in thousands and thousands of pictures would be a daunting task for anyone, the team relied on a machine learning program to help them. People first trained the computer program to pick out things like wall art, plants, books and paint colors in the pictures. And then they handed the task off to the AI program.
“The term for this is transfer learning, but it’s a two-step process,” said Liu. “The first step is to classify the images into categories, such as living rooms, kitchens, bedrooms, and also outdoor space. Then, we use object detection. The program will draw boxes around objects in the rooms, like wall art and books, and then the program counts how many of those objects we have in each image.”
The researchers accessed the photos through the application program interface — or API — of the Airbnb website.
Liu and Andris also worked with Zixuan Huang, graduate assistant in geography, University of Utah and Sohrab Rahimi, doctoral student in architecture, Penn State.
The researchers said that besides the look into living rooms around the world, the work may show scientists new ways that AI can help efforts to explore and understand human cultures.