Jan. 23, 2008 -- When it comes to buying clothes, what looks good is a combination of personal preference and fashion trends. But shoppers can feel overwhelmed by the choices and often seek advice from friends.
Now a dressing room equipped with cameras and interactive displays could lead to intelligent fitting rooms that connect shoppers to a social fashion network, where they can see themselves and others wearing different outfits.
The system can help shoppers compare their dressing room choices before purchase and can suggest alternative styles. The program can also recommend other clothes available on the store's racks.
"The system improves [a shopper's] confidence in the decision-making process and improves their shopping experience," said Wei Zhang, a Ph.D. student in electrical engineering at Oregon State University in Corvallis.
Zhang developed the system with Takashi Matsumoto of Keio University and Juan Liu, Maurice Chu and Bo Begole of PARC in Palo Alto, Calif.
Interactive dressing rooms and social fashion networking are already a part of our society. A Prada store in New York City has a "Magic Mirror" that takes a video image of the shopper, which can then be sent via email or SMS to friends, who can reply with a thumbs up or down.
And the Web site Shareyourlook lets fashionistas upload images of themselves wearing different outfits, which others can then comment on.
But neither system provides a real-time interactive experience.
Zhang's Responsive Mirror uses cameras, displays, machine learning and computer vision software to produce an interactive display and intelligent clothes retrieval.
In the Intelligent Fitting Room, the shopper sees a mirror flanked by displays. Directly in front, she sees herself in the actual mirror. To the left of the mirror, she sees herself on a display wearing the outfit she tried on previously. This allows her to compare multiple outfits at the same time.
To the right of the mirror, the shopper sees another display showing images of other people wearing similar clothes or completely different styles. This is meant to give the shopper information about the social context of her choice and to provide her with alternative fashions she might like to try.
"I haven't seen any other projects that have taken this seriously. I think it's an interesting first step," said Henry Lieberman, a research scientist in the Media Lab at the Massachusetts Institute of Technology in Cambridge. Lieberman's research group developed a fashion recommender, which Discovery reported on last year.
But when it comes to fashion, there are lots of variables, Lieberman said, such as differences between men and women and the colors and sizes each group prefers. Some people might choose baggy clothes, while others might like their clothes tight. Being able to understand which features are important to people will be key to making the system work.
"Obviously there are a lot of technical issues that need to be dealt with one by one before you can have a satisfactory customer experience," said Lieberman.
"It's nice that they're thinking about a situation where technology is not traditionally used," he added.
Zhang's system is owned by PARC and is still in the prototype stage. But it's not limited to just the dressing room. Zhang thinks that the clothes-matching and retrieval technology could be shrunk down to fit mobile device applications.
For example, if a person is window shopping and sees an item in an expensive boutique, she could snap a photo of it with her camera phone and then upload it to a Web site. There, the image could be compared to other clothing in a database, which could provide the user with similar clothing available at a lower price nearby.
Related Links: