Abstract
We present a novel real time multi camera system for tracking 3D position of a face in an office environment. The system uses a combination of low level and high level features for robust tracking of a face in two dimensional (2D) view of each camera. Then the 2D estimates are combined through Quality Threshold clustering to produce an estimate of the 3D position, which is fed back to the 2D trackers to correct any possible errors.
Type
Publication
In IEEE European Conference on Visual Media Production
