[Insight-users] registration of segmented image

[Luis Ibanez]

Hi Orientation f,

Given that you have already segmented the structures out of
the CT and the MRI image, You can perform registration among
two segmented (binary) images by using the MeanSquare image
metric or the MatchCardinality image metric:

http://www.itk.org/Insight/Doxygen/html/classitk_1_1MatchCardinalityImageToImageMetric.html


     Regards,


         Luis


-------------------
orientation f wrote:
> Thanks  for your reply, first.
> Now, I have already segmented the CT and MR image, i.e., extracting the 
> bone from CT image and soft tissue from MR image. Next, what should I do 
> is to fuse them together after registration. So is there any 
> registration methods that are suitable for my work? It would be greatful 
> if you can give me any further hints.
> Best wishes!
> 
> 2008/12/11 Luis Ibanez <luis.ibanez at kitware.com 
> <mailto:luis.ibanez at kitware.com>>
> 
>     Hi Orientation f,
> 
>     You may want to start by using the following combination
> 
> 
>      1) VersorRigig3DTransform
>      2) VersorRigid3DTransformOptimizer
>      3) MattesMutualInformationImageMetric
>      4) LinearInterpolator
> 
> 
>     Note that this is assuming that your are registering the CT image
>     to the MRI image, not the CT segmentation to the MRI segmentation.
> 
> 
>      Regards,
> 
> 
>         Luis
> 
> 
>     --------------------
> 
>     orientation f wrote:
> 
>         Hi,
>         After segmenting the 3D CT and MRI image of knee, I want to fuse
>         them together.
>         But I find there are so many registration methods in ITK
>         framework, which method may give a better result of my work?
>         Please give me some advice.
>         thank you very much.
> 
>