[Insight-users] itkMultiResolutionPDEDeformableRegistration and how to avoid maximum kernel size warning

Luis Ibanez luis.ibanez at kitware.com
Tue May 25 14:11:44 EDT 2010


Hi Maarten,

The warning from the GaussianOperator indicates
that you are asking for a Variance) that is too large
for being accurately approximated by the kernel
size that you are selecting.



Your options are clearly:

A) Reduce the Variance
    operator->SetVariance( )

     or

B) Increase the Kernel size by calling
     operator->SetMaximumKernelWidth( )

    or

C) Settle for a lower precision by calling
     operator->SetMaximumError()


That said,
I'm having trouble finding why is that
you get to use the GaussianOperator
when running the class

MultiResolutionPDEDeformableRegistration

??


Maybe the Warning is comming from the

   MultiResolutionPyramidImageFilter   ?

In which case, you should call the method:


   pyramidFilter->SetMaximumError( ...)



   Regards,


       Luis

---------------------------------------------------------------------
On Wed, May 12, 2010 at 2:26 PM, Maarten Beek <beekmaarten at yahoo.com> wrote:
> Hi all;
>
> I am using the itkMultiResolutionPDEDeformableRegistration class to align
> two images. The images typically have a voxelsize of 0.0027 mm. I would
> subdivide the images in relatively small parts (perferably more than 10
> pyramid levels) because the ultimate goal is the generate an accurate
> displacement and strain field.
>
> Whether I use the MultiResolutionPyramidImageFilter (default in
> itkMultiResolutionPDEDeformableRegistration) or change the pyramids into the
> itkRecursiveMultiResolutionPyramidImageFilter type, I often get a warning
> concerning the kernel width being larger than its maximum width.
>
> I was wondering a) how to avoid this warning (both pyramid types only enable
> to set the maximum error), b) what happens when this warning is given (it
> seems itkGaussianOperator just generates a smaller - more inaccurate-
> kernel; but how would this affect the registration result and does
> continuing to apply pyramid levels make sense after this warning is given?)
>
> I was hoping for another solution than to create new
> itk(Recursive)MultiResolutionPyramidImageFilter classes and include a
> SetMaximumKernelWidth() function
>
> Thanks Maarten
>
>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.html
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>
>


More information about the Insight-users mailing list