The AcceleGlove - Capturing Hand Gestures in Virtual Reality
By Mike Hanlon
June 4, 2004
A glove that translates the hand movements of sign language into written text or speech is just one on the incredible benefits that will flow from developments in VR technology like the AcceleGlove.In constant development since the summer of 2000 the prototype uses a glove system that enables 'Whole Hand Input' using accelerometers attached to a leather glove. The latest design incorporates a two-link arm section to accommodate the recognition of a wider range of gestures.The system captures the four distinctive components of hand gestures -handshape, hand orientation, location, and movement - all measured relative to the position of the users' body.
A combination of the four components measured by the VR can be used to express the one-handed gestures of the American Sign Language (ASL) with the help of a simple algorithm. When tested on a lexicon of 30 signs extracted from different published papers on the subject the system achieved a 98% recognition rate and when the lexicon was expanded to 180 signs, the accuracy achieved was 93%.New signs can be added to the lexicon is performed without modifying the classification algorithm, the same system can be applied to other hand gestures that can be deconstructed into those four features.
Gesture recognition devices like the AcceleGlove have an array of potential applications in areas such as robotics, tele-commanding, silent communication, virtual reality, graphics, design and body kinematics as well as sign language and of course, video games.The Inventors of the AccelGlove, Jose L. Hernandez-Rebollar and Nicholas Kyriakopoulos from the The George Washington University, hope to have a product developed before the end of this year. More on software developed by the Institute for ASL can be found at www.idrt.com.