DocumentCode
3311596
Title
A FEM-based deformable model for the 3D segmentation and tracking of the heart in cardiac MRI
Author
Pham, Q.C. ; Vincent, F. ; Clarysse, P. ; Croisille, P. ; Magnin, I.E.
Author_Institution
CREATIS CNRS, INSA, Villeurbanne, France
fYear
2001
fDate
2001
Firstpage
250
Lastpage
254
Abstract
We present a new approach for the segmentation and tracking of the heart from cardiac MR multi-slice and multi-phase image sequences. An a priori model of the object to be segmented is defined. It is composed of a topology and a geometry of the object, and associated elastic material properties. The a priori model is immersed into the image data and submitted to a force field which pulls the model´s interfaces towards the image edges. The equilibrium of the system is achieved by the minimization of the model´s energy using the finite element method. The active region model (ARM) has been successfully applied to the segmentation and tracking of the heart in cardiac MRI. The deformed model directly provides clinically relevant parameters such as volumes and also physical and local parameters such as strains and stresses
Keywords
biomedical MRI; cardiology; finite element analysis; geometry; image segmentation; image sequences; topology; tracking; 3D heart segmentation; FEM-based deformable model; active region model; cardiac MRI; clinical parameters; elastic material properties; finite element method; force field; heart tracking; image data; local physical parameters; multi-phase image sequences; multi-slice image sequences; object geometry; object topology; Deformable models; Finite element methods; Geometry; Heart; Image segmentation; Image sequences; Magnetic resonance imaging; Material properties; Minimization methods; Topology;
fLanguage
English
Publisher
ieee
Conference_Titel
Image and Signal Processing and Analysis, 2001. ISPA 2001. Proceedings of the 2nd International Symposium on
Conference_Location
Pula
Print_ISBN
953-96769-4-0
Type
conf
DOI
10.1109/ISPA.2001.938636
Filename
938636
Link To Document