A Robust Real-Time Face Tracking Using Head Pose Estimation for a Markerless AR System

Abstract

In this paper we present an extension to the Kinect- Fusion algorithm that allows a robust real-time face tracking. This is achieved altering the original algorithm such that when the tracking algorithm fails, it uses a head pose estimation to give an initial guess to the Iterative Closest Point (ICP) algorithm. We show that this approach can handle more face pose changes and variations than the original KinectFusion's tracking.