Hundreds of workers on Amazon’s Mechanical Turk labor market have been slaving away deciding which clothes look good with what, all to save you the trouble of trying things on.
They’ve been hired by the National University of Singapore to categorize over 25,000 images of clothes, decide which go together, and help create a ‘Magic Closet’. Just tell it the occasion, and it’ll choose the perfect outfit.
The app is – wait for it – “the first one to investigate the task of occasion-oriented clothing recommendation and clothing pairing, which mines the matching rules among more semantic attributes from real images automatically,” explains its creator, Dr Yan Shuicheng.
The Magic Closet first catalogs your own clothes, using a Microsoft Kinect camera to identify features such as sleeve length, colour and collar type. Then, when you ask it for an outfit, it selects and pairs these up with images from online shopping websites and Flickr, and applies them to your image.
“The Magic Closet can be used as an iPhone app or Android app for personal use or as a plug-in system for online shops to help customers choose suitable clothes,” says Yan. “It can also be applied in stores here to let their customers try out different clothes virtually.”
Of course, handing your clothing choices over to a bunch of strangers working on minimum wage might not be for you. But, says Yan, the Magic Closet is nearly ready for commercial release.
“We are currently fine-tuning it to increase its ability to detect users’ shape and figure from various angles besides the frontal view,” he says.
“We also plan to make use of the large-scale annotation data collected from Amazon’s Mechanical Turk, such as ‘red’, ‘V-shaped collar and ‘long sleeve’, and develop some machine learning algorithms to further refine the retrieval performance.”