[Insight-users] Adaptive Meshing for FEM - Registration

Nicholas Tustison ntustison at gmail.com
Mon Jun 12 12:14:58 EDT 2006


As I mentioned, there is no "casting" or conversion operation which  
would give you a FEM mesh from an ITK mesh.  The plan was to  
originally derive the FEM mesh from the ITK mesh but this proved to  
be prohibitively difficult.  Custom meshing is performed by creating  
a mesh using some external program and then feeding that directly  
into the FEM project (e.g. into the itkFEMRegistrationFilter class).   
This is definitely an area of improvement for the FEM library.

As regards the other question --- the current incarnation of the FEM  
library requires that each regularization scheme (e.g. membrane) be  
instantiated in a separate class for each element type.  This is why  
there exists one membrane class for the hexahedral elements as well  
as a separate membrane class for the tetrahedral elements.  Only one  
element type is used (the default is hexahedral).  The other  
declaration is superfluous.

Nick


On Jun 12, 2006, at 11:57 AM, Markus Weigert wrote:

> Hi Nick,
>
> if I see this the right way, the only way to achieve this kind of  
> meshing would
> be to use the ITK - Mesh and create a mesh - input file from this  
> mesh, which the FEM - Filter
> can read, right?
> I didn't have a look at the itk mesh yet, but is it appropriate for  
> tatrahedral mesh generation which
> can be used to create a fem - mesh?
>
> Further question:
>
> What is the concrete function of these lines (taken from software  
> guide), why two different
> membrane - types, which one is actually created?:
>
>  typedef itk::fem::Element3DC0LinearHexahedronMembrane    
> Element3DType;
>  typedef itk::fem::Element3DC0LinearTetrahedronMembrane   
> Element3DType2;
>   {
> //  Software Guide : BeginCodeSnippet
>   Element3DType::LoadImplementationFunctionPointer fp =
>      
> &itk::fem::ImageMetricLoadImplementation<ImageLoadType>::ImplementImag 
> eMetricLoad;
>   DispatcherType::RegisterVisitor((ImageLoadType*)0,fp);
> //  Software Guide : EndCodeSnippet
>   }
>   {
>   Element3DType2::LoadImplementationFunctionPointer fp =
>      
> &itk::fem::ImageMetricLoadImplementation<ImageLoadType>::ImplementImag 
> eMetricLoad;
>   DispatcherType2::RegisterVisitor((ImageLoadType*)0,fp);
>   }
>
>
> Regards,
> Markus
> ----- Original Message -----
> From: Nicholas Tustison
> To: Markus Weigert
> Cc: insight-users at itk.org
> Sent: Monday, June 12, 2006 5:41 PM
> Subject: Re: [Insight-users] Adaptive Meshing for FEM - Registration
>
> Hi Markus,
>
> A straightforward conversion between an ITK mesh and a FEM mesh is  
> non-existent.  I'm reworking the FEM mesh class now and maybe I'll  
> come up with something but until then...
>
> The current FEM framework allows for an external mesh to be brought  
> into the FEM project but that, in addition to the default  
> rectilinear mesh that is created when no external mesh is  
> specified, is the extent of the current FEM library meshing  
> capabilities.
>
> Nick Tustison
>
>
> On Jun 12, 2006, at 11:32 AM, Markus Weigert wrote:
>
>> Hi all,
>>
>> in the insight book ("Insight into Images") from Terry S. Yoo,
>> there is a nice example of adaptive lung meshing (page 331).
>> My question is, how can such a meshing be obtained with ITK in 3D?
>> I suppose, one can use the ITK - Mesh framework for this task,
>> but the ITK Mesh would need to be casted to a FEM - Mesh.
>> Can this casting be done in a straight forward way, or is it more  
>> complicated?
>>
>> Also, this would be interesting because one could start to segment  
>> an image
>> and mesh individual surfaces to model each tissue with different  
>> elasticity.
>> Any idears on this???
>>
>> Regards,
>> Markus
>>
>>
>>
>> _______________________________________________
>> Insight-users mailing list
>> Insight-users at itk.org
>> http://www.itk.org/mailman/listinfo/insight-users
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/insight-users/attachments/20060612/9f69f2e1/attachment.htm


More information about the Insight-users mailing list