Face Reconstruction through Active Stereovision. Reconstruction de Visages par Stéréovision Active

Face Reconstruction through Active Stereovision

Reconstruction de Visages par Stéréovision Active

Régis Vaillant Isabelle Surin 

Thomson-CSF, Laboratoire central de Recherches, Domaine de Corbeville, 91404 Orsay

Groupement Traitement et Simulation TTD, Rue GuynemerBP 55, 78283 Guyancourt Cedex

Corresponding Author Email: 
vaillant@thomson-lcr.fr
Page: 
201-211
|
Received: 
5 May 1994
|
Accepted: 
N/A
|
Published: 
30 April 1995
| Citation

OPEN ACCESS

Abstract: 

The automatic face recognition is a very attractive problem and several solutions can be found in the literature. Most of them rely on the analysis of the images acquired by a classical CCD camera. These images are treated and given us input to a discrimination algorithm. However, the information contained in the image is relatively poor and it is very likely that these techniques will fail in the case of a large database with a lot of people. The surface of the faces is a very discriminant information so in this article, we propose a stereovision algorithm which can be used for the acquisition of the surface offaces. The problem of matching between the pattern and its images is solved using the epipolar constraint and the local coherency constraint. Some experimental results are shown. 

Résumé

La reconnaissance automatique de visages est un problème qui suscite beaucoup d'intérêt et pour lequel divers algorithmes basés sur l'utilisation d'images acquises par des caméras CCD ont été proposés. L'information fournie aux algorithmes de discrimination mis en place dans ce type de solution est assez fruste et ces algorithmes sont peu susceptibles de fonctionner de manière robuste pour des bases comprenant un grand nombre d'individus . La surface du visage doit pouvoir être beaucoup plus discriminante. Dans cet article, nous proposons un algorithme de stéréovision active qui permet l'acquisition de surfaces de visages . Le problème de la mise en correspondance entre les éléments du motif projeté et leur image est résolu en utilisant la contrainte épipolaire et une contrainte de cohérence locale. Des résultats expérimentaux sont présentés. 

Keywords: 

ActiveStereovision, Surface reconstruction, Face.

Mots clés

Stéréovision Active, Reconstruction de surface, Visage.

1. Introduction
2. Le Motif- La Calibration
3. Etiquetage
4. Résultats Expérimentaux
5. Conclusion
Annexe
  References

[1] Nicolas Ayache and Francis Lustman. Fast and Realible Passive Trinocular Stereovision. In First International Conference on Computer Vision, June 1987. 

[2] D. Beymer. Face Recognition under Varying Pose. In Computer and Pattern Recognition, June 1994. 

[3] J.Y Cartoux. Formes dans les images de Profondeur . Application à la Reconnaissance et à l'authentification de Visages. PhD thesis, Université Blaise Pascal de Clermont-Ferrand, October 1989. 

[4] J.Y Cartoux, J.T. Lapreste and M. Richetin. Face Authentification by Profile Extraction from Range Images. In Proceedings of IEEE Workshop on Interpertation of 3D Scenes, pages 194-100, November 1989. 

[5] Stanley M. Dunn, Richard L. Keizer and Jongdaw Yu. Measuring the Area and Volume of the Human Body with Structured Light.IEEE Transactions on Systems Man and Cybernetics, 19 : 1350-1364, November/December 1989. 

[6] OliverFaugeras and Giorgio Toscani. The Calibration Problem for Stereo. In Proceedings of CVPR'86, Miami Beach Florida, pages 15-20, 1986. 

[7] Olivier D. Faugeras. Three-dimensional Computer Vision : a geometric viewpoint. MIT Press, 1993. 

[8] Pascal Fua. Combining Stereo and Mononuclear Information to Compute Dense Depth Maps that Preserves Discontinuties . In12th International Joint Conference on Artificial Interlligence, pages 1292-1298, August 1991. 

[9] Gordon Gaile Gibson. Face Recognition from Depth and Curvature. PhD thesis, Harward University, Cambridge, Massachusetts, September 1991. 

[10] L.D. Harmon, M.K. Khan, Richard Lasch and PF. Ramig. Machine Identification of Human Faces.Pattern Recognition, 13 :2:97-110, 1980. 

[11] GongzhuHu and George Stockman. 3-D Surface Solution Using Structured Light and Constraint Propagation. Pattern Analysis and Machine Intelligence, 11(4) :390-402, April 1989. 

[12] John C. Lee and Evangelos Milios. Matching Range Images of Human Faces. InThird International Conference on Computer Vision, 1990. 

[13] Yael Moses, Yael Adini and Shimon Ullman. Face Recognition : the Problem of Compensating for Changes in Illumination Direction. InEuropean Conference on Computer Vision, volume 1, pages 286-296, Stockholm, May 1994. 

[14] LaurentNajman, Régis Vaillant andEtienne Pernot. From Face Sideviews to Identification. In Gianni Vemazza, editor, Image Processing : Theory and Applications, San Remo, Italy, 1993. Elsevier. [15] A. Pentland, B . Moghaddam, O. Starner, T. Oliyide and M. Turk. ViewBased and Modular Eigenspaces for Face Recognition . In Computer Vision and Pattern Recognition, June 1994. 

[16] Luc Robert andOlivierFaugeras. Curve-Based Stereo : Figural Continuity and Curvature. InComputer Vision and Pattern Recognition, 1991. 

[17] Ashok Samal and Prasana Iyengar. Automatic recognition and analysis of human faces and facial expressions : a survey. Pattern Recognition, 25(1) 65-77, 1992. 

[18] Isabelle Surin.Reconstruction de Visages parstéréovisionActive. Technical Report ASRF-92-5, Thomson-CSF, L.C.R., September 1992. Rapport de stage du DEA Robotiqueet Vision par Ordinateur de l'Université de Nice Sophia-Antipolis. 

[19] Thom. Production AutomatiquedeModèles Numériques Tridimensionnels d'Objets en Temps Quasi-Réel à l'aide d'une caméraCCDVidéo. Research report. IGN, November 1988. 

[20] Giorgio Toscani. Système de Calibration et Perception du Mouvement en vision artificielle. Ph.D thesis, Université de Paris-Sud Centre d'Orsay, December 1987. 

[21] Mathew A. Turk and Alex P. Pentland. Eigenfaces for Recognition.Journal of Cognitive Neuroscience, 3(1) :72-86, 1991. 

[22] YF. Wang and P. Liang. A New Method Computing Intrinsic Surface Properties. InConference on Computer Vision and Pattern Recognition, pages 235-240, 1989.