Tarantula is an Octree-based mesh generator that generates unstructured, boundary fitted, locally refined conformal FE grids. Tarantula is external software developed by Ferdinand Kickinger, but is custom designed for biomedical application in general and for cardiac modeling in particular.
To be ideally suited for cardiac bidomain simulations Tarantula has been designed with the following goals in mind:
  1. Tarantula generates unstructured grids to allow accurate and smooth representations of the organs surfaces. This ensures general applicability of the meshes including de- fibrillation studies since unstructured grids lack jagged boundaries that are typical for structured grids (which introduce spurious currents due to tip effects);
  2. Unstructured grids can be generated adaptively such that the spatial resolution varies. Fine discretizations with little adaptivity can be used to model the myocardium thus minimizing undesired effects of grid granularity on propagation velocity, and coarser elements that grow in size with distance from myocardial surfaces are generated to represent a surrounding volume conductor (e.g., tissue bath or torso). Using adaptive mesh generation techniques facilitates the execution of bidomain simulations with a minimum of overhead due to the discretization of a surrounding volume conductor.
  3. Conformal meshes should be generated to enable the use of standard FE techniques without having to resort to discontinous FE methods to deal with hanging nodes.
  4. The mesh generator should be able to accept segmented image stacks as input to avoid the tedious work of finding tesselated representations of the cardiac surface as required by Delaunay based mesh generators.
  5. The method should be able to adjust the mesh resolution on the fly independently of the resolution of the image stack. Structures which are below a choosen mesh resolution will be ignored.
:: Features
Some of the salient features built into Tarantula are shown here. A high-resolution MRI image stack (Imaging data are courtesy of Dr. P. Kohl et al) of a rabbit left ventricular wedge preparation is used directly as input for Tarantula. The wedge is immersed in a tissue bath of slab-like geometry. As can be seen, the bath is adaptively discretized with using increasingly coarser elements with distance from the wedge. The wedge itself is discretized at an average resolution of 75 Ám. Assuming that the MRI stack is available in a segmented form, it takes about 2 minutes to generate this mesh in a fully automatic fashion. The same mesh generation run can be repeated with to generate a mesh of coarser resolution. As can be seen, finer structures related to cleft spaces and vascularization have disappeared.
MRI Segmented -> Adaptive Hex-dominant Hybrid Mesh
<< click on the image to enlarge!
<< click on the image to enlarge!
Figure 11: A MRI image stack is feed into tarantula to generate a mesh representing a rabbit left ventricular wedge preparation. The surrounding bath is discretized using increasingly coarser elements with distance from the myocardial surfaces.
[TARANTULA] . previous page . next page .