'Magic Closet' helps people decide what to wear
Scientists at the Chinese Academy of Science and the National University of Singapore have developed a clothing-selection system that help people decide what to wear by suggesting clothing from their own wardrobes based on the event they plan to attend.
The system, called Magic Closet, first catalogues clothes in a wardrobe, using a Microsoft Kinect camera to identify defining features such as sleeve length, colour and collar type, New Scientist reported.
Then, when user ask it for an outfit, the software selects the best matches from its database of 25,000 outfit images, skimmed from online shopping websites and photo-sharing site Flickr - and presents pictures of its selected items of the clothing.
Those database images are then fed into Amazon's Mechanical Turk labour market, where online workers categorised the outfits by attributes like collar type, it said.
They also determined if the pictured clothes were more suitable for some occasions - a first date, say - than others, such as a wedding.
An attribute or occasion was deemed legitimate when more than half the workers agreed.
In addition to helping people decide what to wear, the system can also help online shoppers find clothes that go with what they already own.