Optical tracking and calibration of tangible interaction devices

Abstract

In this paper, a novel optical tracking and object calibration system is presented for the recognition and pose estimation of tangible interaction devices for virtual and augmented reality systems. The calibration system allows a user to automatically generate models of the relative positions of point-shaped markers attached to interaction devices, simply by moving them in front of the cameras. There are virtually no constraints on the shape of interaction devices. The tracking method takes the calibrated models as input, and recognizes devices by subgraph matching. Both the calibration and tracking methods can handle partial occlusion. Results show the proposed techniques are efficient, accurate, and robust.