Finite Element Meshing for Cardiac Analysis

Yongjie Zhang, Chandrajit Bajaj


The Center for Computational Visualization (CCV)
Institute for Computational Engineering and Sciences & Dept. of Computer Sciences
The University of Texas at Austin

This application paper presents details of the technique we developed to produce an adaptive and quality tetrahedral finite element mesh model of a human heart. Beginning from a polygonal surface model consisting of twenty-two components, we first edit and convert it to volumetric gridded data. A component index for each cell edge and grid point is computed for assisting the boundary and material layer detection. Next we extract adaptive and quality tetrahedral meshes from the volumetric gridded data using our Level Set Boundary and Interior-Exterior (LBIE) Mesher. The mesh adaptivity is controlled using a feature sensitive error function. Multiple layers with different materials were identified and meshed. Furthermore, one of the heart valves in the input multi-component surface model was replaced. The extracted final tetrahedral mesh is being utilized in the analysis of cardiac fluid dynamics via finite element simulations.

Paper Download

Finite Element Meshing for Cardiac Analysis (pdf) , ICES Technical Reoprt 04-26, the University of Texas at Austin, 2004.


Links

  • A website communicating with our collaborators

  • Project description on our CCV website


    Results

    (Each image is linked to a higher resolution image.)
    1. The original heart model from NYU* and the modified model.


    (a) - the aortic valve; (b) - the tricuspid valve; (c) - the pulmonary valve; (d) - the mitral valve; (a') (b') (c') and (d') - modified valves; (e) - the `foramen ovale' connecting the left and right atriums. The original (e) and modified (e) are compared in the bottom row. Note*: With permission of New York University, © Copyright 1994-2004.


    2. Adaptive tetrahedral meshes for the heart model.


    (a) - the heart model viewed from outside; (b) - the result of boundary detection in wireframe, each of the twenty-two components of the heart model is represented by a different color, the relationship between the color and heart components is listed in Figure \ref{nodal}; (c) - a cross section of the adaptive tetrahedral mesh, it is obvious that the valves have the finest mesh, thin structures are identified by the feature sensitive error function, and adaptive meshes are generated to preserve correct topology.


    3. Four valves with gaps in the adaptive tetrahedral mesh of the cardiac model.


    (a) - the aortic valve; (b) - the tricuspid valve; (c) - the pulmonary valve; (d) - the mitral valve.


    4. The result of material layer detection in the extracted tetrahedral meshes -- the interface between the mitral valve and the myocardium is identified.


    (a) - a cross section of the mitral valve before material layer detection; (b) - a cross section of the mitral valve after material layer detection.