Face 3D Pose Estimation Using a Generic 3D Face Model and Facial Feature Extraction

by Barmpoutis, A. | Nikolaides, N. | Pittas, I.

In Proceedings of PCI2003: Panhellenic Conference in Informatics, November 21-23, 2003, pp. 519-526.

Category:

Description

In this paper an algorithm that processes a video sequence for humean face 3D pose estimation, is presented. The procedure that is followed is described briefly below. In the beginning, facial features are extracted for each frame. Afterwards, these are used in order to make an initial estimation of their position in 3D space. The results produced are optimized, either by taking into consideration anthropometric features, or by using a 3D model of the human face. In that way, the initial prediction is greately improved and the resulting accuracy is more than satisfactory. Such rechniques could process videos displaying news, journalists, actors or even people in general, or even be used in object-based techniques for video coding (eg. MPEG-4), in machine vision applications and in human computer interaction enviroments.

Additional information

Author

Barmpoutis, A., Nikolaides, N., Pittas, I.

Journal

In Proceedings of PCI2003: Panhellenic Conference in Informatics

Year

2003

Month

November 21-23

Pages

519-526

URL

http://ikee.lib.auth.gr/record/201236?

Citation

Citation

Barmpoutis, A., Nikolaides, N. and Pittas, I., 2003. Face 3D Pose Estimation Using a Generic 3D Face Model and Facial Feature Extraction. In Proceedings of PCI2003: Panhellenic Conference in Informatics, pp. 519-526.

BibTex

@article{digitalWorlds:178,
author = {Barmpoutis, A. and Nikolaides, N. and Pittas, I.},
title = {Face 3D Pose Estimation Using a Generic 3D Face Model and Facial Feature Extraction},
journal = {In Proceedings of PCI2003: Panhellenic Conference in Informatics},
month = {November 21-23},
year = {2003},
pages = {519-526}
}