Untitled Document

3D Face Analysis and Recognition RECOVIS3D

Partners

Institut TELECOM, TELECOM Lille1, CNRS, Floride State University

Partcipants

Mohamed DAOUDI , Professor, Institut TELECOM ; TELECOM Lille 1 (mohamed.daoudi at telecom-lille1 dot eu, daoudi at lifl dot fr), responsable.

Chafik Samir, PhD, INT/USTL (chafik.samir at telecom-lille1 dot eu)

Anuj Srivastava, Professor, Floride State University (USA), anuj at stat.fsu dotedu

Ballihi Lahoucine, PhD Student

There is an increasing interest in analyzing shapes of facial surfaces with many applications including biometrics, facial surgery, video communications, and 3D animation. This interest is fuelled by the advent of cheaper and lighter scanners that can provide high-resolution measurements of both geometry and texture of human facial surfaces. One general goal here is to develop computational tools for analyzing 3D face data. In particular, one is interested in comparing the shapes of facial surfaces. Such a tool can be used to recognize human beings according to their facial shapes, to measure changes in a facial shape due to a surgery, or to study/capture the variations in facial shapes during conversations and expressions of emotions. Additionally, a subproblem may be to find an optimal deformation of one surface into another, under some chosen criterion. These deformations can be useful in defining summary statistics of a collection of faces.

In this project, we have presented a Riemannian framework for analyzing shapes of facial surfaces for the purposes of matching, comparing, and deforming them.In this project we have introduced a new computational framework for analyzing shapes of facial surfaces. The basic idea is to impose a specific coordinate system on facial surfaces using level curves of the surface distance function (measured from the tip of nose). Then, each surface can be represented as a path on the space of closed curves in R3. In figure 1, we show a geodesic path between two facial surfaces of the same person. The main tool presented that we have proposed in this project is the construction of geodesic paths between arbitrary two facial surfaces in the aforementioned set. There are multiple uses of this geodesic construction. The length of a geodesic between any two facial surfaces quantifies differences in their shapes; it also provides an optimal deformation from one to the other. Using Riemannian structure one can define simple statistics such as the sample mean.

Figure 1 : Geodesic Paths between: same person under different facial expressions,

viewed from different viewpoints

A future application of this framework is in biometrics, for example in recognition of humans using shapes of their facial surfaces. This idea will be developed under the project FAR3D accepted by Agence National de la Recherche.

Publications:

Journals

1. A. Srivastava, C. Samir, S. Joshi, and M. Daoudi, Elastic Shape Models for Face Analysis Using Curvilinear Coordinates, Journal of Mathematical Imaging and Vision, downolad (Special issue), accepted (to appear 2008).

2. C. Samir, A. Srivastava, and M. Daoudi, 3D Face Recognition Using Shapes of Facial Curves IEEE Transactions Pattern Analysis and Machine Intelligence, Vol. (28), Issue (11), Page 1858- 1863, Nov. 2006.

3. C. Samir, A. Srivastava, M. Daoudi, E. Klassen, An Intrinsic Framework for Analysis of Facial Surfaces, Submitted to International Journal on Computer Vision (IJCV) march 2007.

International Conferences

4. Samir, M. Daoudi and A. Srivastava, A Framework of Calculus on Facial Surfaces 14th International Conference on Image Analysis and Processing 10-14 september 2007, Modena Italy.2007.

5. C. Samir, M. Daoudi and A.Srivastava Human Identification Using Facial Curves with Extensions to Joint Shape-Texture Analysis, 2nd International Conference on Computer Vision Theory and Applications, 8 - 11 March, 2007 Barcelona, Spain.

6 C. Samir, A. Srivastava, M. Daoudi, 3D Face Recognition Using Shapes of Facial Curves, Statistical inferences on nonlinear manifolds with applications in signal and image processing, special session IEEE ICASSP 2006, May 14-19, Toulouse France.

National Conferences

7. M. Daoudi, A. Sriavastava and C. Samir, Analyse des Variations des Formes des Surfaces Faciales Utilisant la Géométrie Riemannienne, in 16e congrès francophone AFRIF-AFIA Reconnaissance des Formes et Intelligence Artificielle (RFIA 2008). Amiens, France, Jan. 21-25 2008, download

8. Chafik Samir, Mohamed Daoudi, et Anuj Srivastava Analyse Riemannienne des Variations des Formes Faciales, accepté journées d'étude et d'échange COmpression et REprésentation des Signaux Audiovisuels Montpellier, 8-9 novembre 2007.

9. C. Samir, M. Daoudi, et A. Srivastava, Fusion de la Texture et de la Géométrie pour la Reconnaissance de Visage 3D Traitement et Analyse de l'Information : Méthodes et Applications Hammamet, Tunisie — 23 au 26 mai, special session Reconnaissance faciale 2D/3D (invited paper).

10. C. Samir, M. Daoudi, A. Srivastava, Reconnaissance de Visages 3D Utilisant l'Analyse de Formes des Courbes Faciales, 10èmes Journées CORESA (COmpression et REprésentation des Signaux Audiovisuels), Caen, France, 9-10 Novembre 2006. (The best PhD student paper award at CORESA, 2006). Workshops

11. Poster presented by Chafik Samir in Geometry and Statistics of Shape Spaces (workshop), organized by David Mumford and Laurent Younes, and SAMSI (Statistical and Applied Mathematical Sciences Institute), North Carolina, USA, July 2007.

12. C. Samir, M. Daoudi and A. Srivastava, A Framework of Calculus on Facial Surfaces, International Workshop of the EU Network of Excellence DELOS and MUSCLE on VISUAL AND MULTIMEDIA DIGITAL LIBRARIES (VMDL07) (invited paper), 13-14 september 2007, Modena Italy.