Estimation combinée de forme et de mouvement par champs markoviens: application à l'imagerie cardiaque scanner multibarrette
Joint Shape and Motion Estimation using Markovian Fields : Application to Multislice Computed Tomography Cardiac Imaging
OPEN ACCESS
We propose a method for joint surface and non-rigid motion estimation from three-dimensional dynamic sequences. Based on a surface-volume matching, it provides, from one first segmented surface, both motion and deformations of the object of interest along the whole sequence. A Markovian model, combined with a simulated annealing process, estimates the correspondences between the nodes of the surface mesh modeling the object of interest at one time and the voxels of the volume representing the object at the following time. The method has been applied to cardiac surface and motion extraction in Multislice Computed Tomography. Tests realized with simulated motion and on real data have provided promising results.
Résumé
Une méthode d'estimation conjointe de forme et de mouvement non rigide à partir de séquences temporelles tridimensionnelles est proposée. Reposant sur une mise en correspondance surface-volume, elle permet, à partir d'une première segmentation de l'objet d'intérêt, d'estimer le mouvement de l'objet et ses déformations sur toute la séquence temporelle d'observation. Une modélisation markovienne combinée à un algorithme de recuit simulé estime les correspondances entre les nœuds du maillage de surface modélisant l'objet à un instant et les voxels du volume représentant l'objet à l'instant suivant. La méthode a été appliquée à l'extraction de formes et de mouvements cardiaques en tomodensitométrie multibarrette. Les tests, réalisés à la fois avec des mouvements simulés et sur des données réelles, ont donné des résultats prometteurs.
Motion estimation, 3D dynamic CT imaging, cardiac motion, Markov random field
Mots clés
Estimation de mouvement, imagerie scanner tridimensionnelle dynamique, mouvement cardiaque, champ de Markov
[Amini et Duncan 1992] AMINI A. A. et DUNCAN J. S. (1992). Bending and stretching models for lv wall motion analysis from curves and surfaces. Image and Vision Computing, 10(6):418-430.
[Buckberg et al. 2006] BUCKBERG G. D., MAHAJAN A., JUNG, B., MARKL M., HENNIG J. et BALLESTER-RODES M. (2006). MRI myocardial motion and fiber tracking: a confirmation of knowledge from different imaging modalities. European Journal of CardioThoracic Surgery, 29 Suppl 1: S165-S177.
[Chen et al. 1995] CHEN C. W., LUO J., PARKER K. J. et HUANG T. S. (1995). CT volumetric data-based left ventricle motion estimation: an integrated approach. Computerized Medical Imaging and Graphics, 19(1): 85-100.
[Corpetti et al. 2002] CORPETTI T., MÉMIN E. et PÉREZ P. (2002). Dense estimation of fluid flows. IEEE Transactions on Pattern Analysis and Machine Intelligence, 24(3): 365-380.
[Delingette 1999] DELINGETTE H. (1999). General object reconstruction based on simplex meshes. International Journal of Computer Vision, 32(2): 111-146.
[Feldmar et Ayache 1996] FELDMAR J. et AYACHE N. (1996). Rigid, affine and locally affine registration of free-form surfaces. International Journal of Computer Vision, 18(2): 99-120.
[Frangi et al. 2001] FRANGI A. F., NIESSEN W. J. et VIERGEVER M. A. (2001). Three-dimensional modeling for functional analysis of cardiac images: a review. IEEE Transactions on Medical Imaging, 20(1): 2-25.
[Garreau et al. 2004] GARREAU M., SIMON A., BOULMIER D. et GUILLAUME H. (2004). Cardiac motion extraction in multislice computed tomography by using a 3D hierarchical surface matching process. In Proc. IEEE Computers in Cardiology (CinC'04), pages 549-552, Chicago, USA.
[Gorce et al. 1997] GORCE J. M., FRIBOULET D. et MAGNIN I. E. (1997). Estimation of three-dimensional cardiac velocity fields: assessment of a differential method and application to three-dimensional CT data. Medical Image Analysis, 1(3):245-261.
[Guillaume et Garreau 2003] GUILLAUME H. et GARREAU M. (2003). Segmentation de cavités cardiaques en imagerie scanner multibarettes. In 12ème Forum des Jeunes Chercheurs en Génie Biologique et Médical (2003), pages 92-93, Nantes, France.
[Gupta et Prince 1995] GUPTA S. N. et PRINCE J. L. (1995). On Variable Brightness Optical Flow for Tagged MRI. In 14th International Conference on Information Processing in Medical Imaging, pages 323-334.
[Haigron et al. 1998] HAIGRON P., LEFAIX G., RIOT X., et COLLOREC R. (1998). Application of spherical harmonics to the modeling of ana tomical shapes. Journal of Computing and Information Technology, 6: 449-461.
[Heitz et Bouthemy 1993] HEITZ F. et BOUTHEMY P. (1993). Multimodal estimation of discontinuous optical flow using markov random fields. IEEE Transactions on Pattern Analysis and Machine Intelligence, 15:1 217-1232.
[Horn et Schunck 1981] HORN B. K. P. et SCHUNCK B. G. (1981). Determining optical flow. Artificial Intelligence, 17(1-3): 185-203.
[Kambhamettu et al. 2003] KAMBHAMETTU C., GOLDGOF D., HE M. et LASKOV P. (2003), 3D nonrigid motion analysis under small deformations. Image and Vision Computing, 21(3): 229-245.
[Kass et al. 1988] KASS M., WITKIN A., et TERZOPOULOS D. (1988). Snakes: Active contour models. International Journal of Computer Vision, 1(4): 321-331.
[Kervrann et Heitz 1998] KERVRANN C. et HEITZ F. (1998). A hierarchical markov modeling approach for the segmentation and tracking of deformable shapes. Graphical Models and Image Procesing, 60(4):173-195.
[Li 1995] LI S. Z. (1995). Markov random field modeling in computer vision. Springer Computer Science Workbench Series.
[Lorensen et Cline 1987] LORENSEN W. E. et CLINE H. E. (1987). Marching cubes: A high resolution 3D surface reconstruction algorithm. Computer Graphics, 21: 163-169.
[McEachen et Duncan 1997] MCEACHEN J. C. et DUNCAN J. S. (1997). Shape-based tracking of left ventricular wall motion. IEEE Transactions on Medical Imaging, 16(3): 270-283.
[McInerney et Terzopoulos 1995] MCINERNEY T. et TERZOPOULOS D. (1995). A dynamic finite element surface model for segmentation and tracking in multidimensional medical images with application to cardiac 4D image analysis. Computerized Medical Imaging and Graphics, 19(1): 69-83.
[Montagnat et Delingette 2005] MONTAGNAT J. et DELINGETTE H. (2005). 4D deformable models with temporal constraints: application to 4D cardiac image segmentation. Medical Image Analysis, 9(1): 87-100.
[Nastar et Ayache 1996] NASTAR C. et AYACHE N. (1996). Frequencybased nonrigid motion analysis: Application to four dimensional medical images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 18(11): 1067-1079.
[Osher et Sethian 1988] OSHER S. et SETHIAN J. A. (1988). Fronts propagating with curvature-dependent speed : Algorithms based on Hamilton-Jacobi formulations. Journal of Computational Physics, 79:12-49.
[Paragios 2003] PARAGIOS N. (2003). Shape-based segmentation and tracking in cardiac image analysis. IEEE Transactions on Medical Image Analysis, 22:402-407.
[Park et al. 1996] PARK J., METAXAS D., YOUNG A., et AXEL L. (1996). Deformable models with parameter functions for cardiac motion analysis from tagged MRI data. IEEE Transactions on Medical Imaging, 15(3): 290-298.
[Prince et McVeigh 1992] PRINCE J. L. et MCVEIGH E. R. (1992). Motion estimation from tagged MR image sequences. IEEE Transactions on Medical Imaging, 11(2): 571-583.
[Pérez 1998] PÉREZ P. (1998). Markov random fields and images. CWI Quarterly, 11(4): 413-437.
[Ruan et al. 1994] RUAN S., BRUNO A., et COATRIEUX J.-L. (1994). Threedimensional motion and reconstruction of coronary arteries from biplane cineangiography. Image and Vision Computing, 12(10): 683-689.
[Shi et Liu 2003] SHI P. et LIU H. (2003). Stochastic finite element framework for simultaneous estimation of cardiac kinematic functions and material parameters. Medical Image Analysis, 7(4): 445-464.
[Song et Leahy 1991] SONG S. M. et LEAHY R. M. (1991). Computation of 3-D velocity fields from 3-D cine CT images of the human heart. IEEE Transactions on Medical Imaging, 10(3): 295-306.
[Tu et al. 1995] TU H. K., MATHENY A., GOLDGOF D. B., et BUNKEH. (1995). Left ventricular boundary detection from spatiotemporal volumetric computed tomography images. Computerized Medical Imaging and Graphics, 19(1): 27-46