Robust finger tracking with multiple cameras

This paper gives an overview of a system for robustly tracking the 3D position and orientation of a finger using a few closely spaced cameras. Accurate results are obtained by combining features of stereo range images and color images. This work also provides a design framework for combining multiple sources of information, including stereo range images, color segmentations, shape information and various constraints. This information is used in robust model fitting techniques to track highly over-constrained models of deformable objects: fingers