Gesture-Based Computing on the Cheap From: MIT News - 05/20/2010 By: Larry Hardesty Massachusetts Institute of Technology researchers have develop a user interface system that can translate gestures made with a gloved hand into the corresponding gestures of a three-dimensional model of the hand on screen, with almost no lag time. The system uses a standard webcam and an inexpensive multicolored Lycra glove. The glove is covered with 20 irregularly shaped patches that use 10 different colors. The arrangement and shapes of the patches were chosen so that the front and back of the hand would be distinct but also so that similar-colored patches would rarely collide with each other. The other key to the system is an algorithm that can rapidly look up visual data in a database. Once a webcam has captured an image of the glove, the software crops out the background, so that the glove alone is superimposed upon a white background. The software then reduces the resolution of the cropped image to 40 pixels by 40 pixels. Next, it searches through a database containing myriad 40-by-40 digital models of a hand, wearing the distinctive glove, in a range of different positions to find a match in a fraction of a second. Read the entire article at: http://web.mit.edu/newsoffice/2010/gesture-computing-0520.html Links: Robert Wang http://people.csail.mit.edu/rywang/ Computer Science and Artificial Intelligence Laboratory http://www.csail.mit.edu/ Real-time hand-tracking with a color glove http://people.csail.mit.edu/rywang/hand/ -- Thumbs Up for Gesture-Based Computing From: New Scientist - 06/09/2010 By: Shanta Barley A gesture-based computing system with a simple hardware interface has been developed by Robert Wang at the Massachusetts Institute of Technology's Artificial Intelligence Lab. The system is composed of a pair of multi-hued latex gloves, a laptop, and a webcam. Twenty patches of 10 unique colors cover the gloves, which are configured to maintain the best possible separation of hues. When the webcam is employed to track a glove-clad hand, the system can recognize the location of each finger and tell the difference between the front and back of the hand. Upon calculating the hand's position, the system mines a database of 100,000 images of gloved hands in various positions, and then displays the closest match on-screen. The system can replicate gestures in real time by repeating the process several times per second. However, eliminating the gloves is thought by some to be key to gesture-based computing's mass appeal, and the possible viability of markerless motion-capture was recently demonstrated by researchers from Sweden's Royal Institute of Technology. Their system attempts to identify an unclad hand in a stream of video from a webcam by spotting flesh tones, using a reference database containing images of hands picking up various objects. Read the entire article at: http://www.newscientist.com/article/mg20627635.100-thumbs-up-for-gesturebased-computing.html