[Insight-users] Question about
MultiResolutionImageRegistrationMethod
Rosario Sance
rsance at die.upm.es
Wed Dec 13 12:29:42 EST 2006
Hi all,
I'm trying to modify the
MultiResolutionImageRegistration Method in order
to improve the results of a global algorithm containing it.
What I wanted to do is to replace the
DiscreteGaussianFiltering of the smoothing step
of every level in the pyramid by the
AnisotropicDiffusionImageFiltering, so that the
sharp boundaries in the image that help to
distinguish the anatomical structures would be preserved...
For doing so, I have tried to modify:
itkMYMultiResolutionImageRegistrationMethod.txx
itkMYMultiResolutionPyramidImageFilter.h
itkMYRecursiveMultiResolutionPyramidImageFilter.txx
itkMYMultiResolutionImageRegistrationMethod.h
itkMYMultiResolutionPyramidImageFilter.txx
itkMYRecursiveMultiResolutionPyramidImageFilter.h
The problem is that I could change the smoothing filter both in
void
MYRecursiveMultiResolutionPyramidImageFilter<TInputImage,
TOutputImage> ::GenerateData()
(in itkMYRecursiveMultiResolutionPyramidImageFilter.txx )
void
MYMultiResolutionPyramidImageFilter<TInputImage, TOutputImage> ::GenerateData()
(in itkMYMultiResolutionPyramidImageFilter.txx )
so that I included properly the
typedef
GradientAnisotropicDiffusionImageFilter<TOutputImage,
TOutputImage> SmootherType;
and changed the different parameters.
But then I found that the methods
void
MYMultiResolutionPyramidImageFilter<TInputImage,
TOutputImage>::GenerateInputRequestedRegion()
void
MYRecursiveMultiResolutionPyramidImageFilter<TInputImage,
TOutputImage>::GenerateOutputRequestedRegion(DataObject * ptr )
void
MYRecursiveMultiResolutionPyramidImageFilter<TInputImage,
TOutputImage>::GenerateInputRequestedRegion()
are using, to compute requirements for the
smoothing part, the kernel implementation:
typedef GaussianOperator<OutputPixelType,ImageDimension> OperatorType;
How could I also replace this operator by an
anisotropic one? I couldn't find that implementation...
Maybe I'm not getting how does it work, and the
kernel implemented in the GaussianOperator for
the smoothing step is also valid for the
anisotropic case or just I shouldn't need it, but I don't think so...
Would somebody help me?
Thank you very much in advance.
Rosario
Rosario Sance Garzón
Dpto. Ingeniería Electrónica
E.T.S.I. Telecomunicación
Avda. Ciudad Universitaria s/n
E-28040 - Madrid (SPAIN)
Teléfono: +34 91 549 57 00 (Ext. 4220)
Fax: +34 91 336 73 23
___________________________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/insight-users/attachments/20061213/185cd3a3/attachment.html
More information about the Insight-users
mailing list