Robust 3D Head Tracking by Online Feature Registration - Robotics Institute Carnegie Mellon University

Robust 3D Head Tracking by Online Feature Registration

Jun-Su Jang and Takeo Kanade
Conference Paper, Proceedings of IEEE International Conference on Automatic Face and Gesture Recognition (FG '08), September, 2008

Abstract

This paper presents a robust method for tracking the position and orientation of a head in videos. The proposed method can overcome occlusions and divergence problems. We introduce an online registration technique to detect and register feature point of the head while tracking. A set of point features is registered and updated for each reference pose serving a multi-view head detector. The online feature registration rectifies error accumulation and provides fast recovery after occlusion has ended, while preventing divergence problem which frequently occurs in conventional frame-to-frame tracking methods. The robustness of the proposed tracker is experimentally shown with video sequences that include occlusions and large pose variations.

BibTeX

@conference{Jang-2008-10090,
author = {Jun-Su Jang and Takeo Kanade},
title = {Robust 3D Head Tracking by Online Feature Registration},
booktitle = {Proceedings of IEEE International Conference on Automatic Face and Gesture Recognition (FG '08)},
year = {2008},
month = {September},
}