[Insight-users] Cast & Deformable Filter

[Ting Chen]

Hi! Luis and Pieter,
I will check in a new deformable filter, which takes arbitary form of the
input mesh and only use the gradient information to make the deformation.
The balloon force part will be moved to itkBalloonForceFilter.

i have to finish some local tests of the filter before I check it in.

ting

-----Original Message-----
From: Luis Ibanez <luis.ibanez@kitware.com>
To: pieterv@nat-239.azn.nl <pieterv@nat-239.azn.nl>;
insight-users@public.kitware.com <insight-users@public.kitware.com>
Date: Monday, April 29, 2002 3:12 PM
Subject: [Insight-users] Cast & Deformable Filter


>
>Hi Pieter,
>
>I have been trying to reproduce the
>problem that you reported concerning
>the interaction between the itkCastImageFilter
>and the DeformableMesh3DFilter.
>
>Sorry that it took some time to start doing this.
>
> From the code in your message I reconstructed
>the attached source file which compiles without
>problem.
>
>However it is seg. faulting at write.
>
>Could you please look at the parameters
>used in this file and tune them to values that
>you consider appropiated.
>
>I doubt that the problem is not comming from
>the itkCastImageFilter. This is one of the simplest
>filters in the toolkit and use the same structure of
>filters like "add" and "multiply".
>
>The only suspicious thing I found in your code
>is that you are calling "SetOutputImage()" on
>the Deformable filter.  This call is deprecated.
>
>ITK filters allocate and produce their output.
>
>You should only to get the image from the filter.
>
>The fact that your code is running one iteration
>before producing nan's could lead to think that
>memory is being corrupted during the execution
>of the filter.
>
>Could you please take a look at the attached file
>and verify that it reflects correctly the set of operations
>that you are doing in your code. If you can, please
>select appropiated valued for the "const" parameters
>at the top of  main() and send it back to me.
>
>Once we reproduce the problem, we can look for
>the reasons behind and hopefully find a solution.
>
>
>
>  Thanks
>
>
>     Luis
>
>
>
>
>
>