Somewhere in your closet there's a pair of gloves straight out of "Breakin' 2: Electric Boogaloo." Well, if you dust them off and fire up your Webcam, you'll have the beginnings of a nifty gestural interface system, thanks to research at MIT.
Robert Wang of MIT's Computer Science and Artificial Intelligence Laboratory designed a gestural system that tracks a pair of rainbow-colored Lycra gloves to be used with a standard Webcam. The gloves, which cost only a dollar to manufacture, can be used to manipulate virtual objects such as blocks or even complex machinery models (see the proof-of-concept vid below).
The system was presented at the Siggraph 2009 conference, and has gone through several design stages. The current gloves are patterned with 20 shapes and 10 colors to enable sophisticated tracking. The system now calibrates faster and works better in bright-light environments.
The tracker is one of many rival prototypes aiming at "Minority Report"-style gestural control, but it requires very little hardware. It recalls a similar MIT prototype, Sixth Sense, which tracks colored material attached to a user's fingertips.
Wang's system, though, translates the entire hand and fingers into a 3D hand with minimal lag time, achieved through use of an algorithm that matches a low-res image of the gloves to many digital models in a database several hundred megabytes large.
The hand tracker could be used as a video game controller to manipulate virtual objects or weapons, or as an engineering and design tool for digital schematics, according to Wang and colleagues.
As the glove tracker is working well, Wang is looking at developing "similarly patterned shirts" as a relatively cheap and simple method of doing whole-body motion capture.
Lycra gloves and rainbow shirts. Break out the hairspray, shoulder pads, and giant cell phone!