This paper is focused on using headtracking for determination of the 3D orientation of head from 2D frame as a supplement lip-reading method. The task is proposed for determination of position and rotation of the drivers (speakers) head during the speech. We divided the article to three parts. First we describe a module of automatic determination of the head shape in the image. The second part contains the localization of important points in the area of head. Finally we present the determination of the 3D orientation and scale and compare the results of our algorithm and a real measurement.
Cite as: Cisar, P., Zelezny, M. (2004) Detection of face position and 3d orientation in 2d image. Proc. 9th Conference on Speech and Computer (SPECOM 2004), 52-56
@inproceedings{cisar04_specom,
author={Petr Cisar and Milos Zelezny},
title={{Detection of face position and 3d orientation in 2d image}},
year=2004,
booktitle={Proc. 9th Conference on Speech and Computer (SPECOM 2004)},
pages={52--56}
}