New technology to improve the accurately estimating the direction of finger pointing in a 3D space. Real-time, robust tracking of the hand without using marks or gloves.
Hand Pointing Estimation for Human-Computer Interaction
Hand pointing is a classical intuitive gesture for human interaction with computers. Big challenges are still encountered in accurately estimating the direction of finger pointing in a 3D space.
The present technology re-invents hand pointing estimation as a cursor system combining two regular cameras hand region detection, hand finger estimation, feature detection, and 3D pointing direction estimation. An Active Appearance Model (AAM)-based approach is applied to detect and track 14 feature points along the hand contour from the top and side views enabling use for intuitive drawing in 3D space.
Real-time, robust tracking of the hand without using marks or gloves
Works over long distances
The hand pointing estimation for computer interaction can be used to resolve ambiguities inherent in verbal communications, thus opening up the possibility of humans interacting or communicating intuitively with computers or robots by indicating objects or pointed locations either in the 3D space or on the screen.
Intellectual Property Position:
U.S. Patent No. 8,971,572 and patent pending.