In this video clip, we show our prototype system which supports the bear hand interaction with virtual contents.
Our software consists of multiple target tracker, hand tracker, and gesture detector. Multiple target tracker makes possible to track the camera pose robustly with a large change of view or occluded situation. Hand tracker estimates the relative position from camera. Gesture detection specifies the interaction type with virtual contents.
In this approach, the user can interact with virtual contents (i.e. virtual globe in this video) by his/her bear hand.