[Insight-users] Re: Landmark Based transform initializer
Luis Ibanez
luis.ibanez at kitware.com
Tue May 24 11:16:37 EDT 2005
Hi Kevin,
The direction of the Transforms in ITK is a common
source of confusion for new users of the image
Registration Framework
We have discussed this topic in many occasions and
we have captured a summary of the rationale behind
the direction of the ITK Transforms in the ITK
Software Guide
http://www.itk.org/ItkSoftwareGuide.pdf
In particular you will find useful to look at the
following sections:
1) "Geometric Transformations"
Section 6.9
pdf-page 245
2) "Direction of the Transform Mapping"
Section 8.3.1
pdf-page 318
In summary the reasons are:
A) The purpose of registration is to resample
the *Moving* image into the coordinate system
of the *Fixed* Image
B) This resampling requires to visit all the points
in the *Fixed* image and to figure out what
intensities from the *Moving* image should be
assigned to them.
Please read the Software Guide sections indicated
above, and let us know if you still have any
further questions or concerns.
Note, of course,
that if you *really* need the Inverse Transform,
you could then just swap the Fixed and Moving
Landmarks in order to suit your needs.
Or, you can also get the current Transform,
and invoke its GetInverse() method.
Regards,
Luis
-------------------------
Kevin H. Hobbs wrote:
> On Mon, 2005-05-23 at 06:49 -0400, Karthik Krishnan wrote:
>
>>Kevin:
>>
>>The class supports the Rigid2DTransform as well as the
>>VersorRigid3DTransform now. Please update your cvs repository. For usage,
>>please see the test
>>Testing/Code/Common/itkLandmarkBasedTransformInitializerTest.cxx
>>
>>Thanks
>>Regards
>>Karthik
>
>
>
> I've put up a bit of code using the new
> itkLandmarkBasedTransformInitializer.
>
> http://crab-lab.zool.ohiou.edu/kevin/Slice_Landmark_Registration/itk_landmark_transform.cxx
>
> It's mostly cobbled together from
> Insight/Examples/Filtering/ResampleImageFilter6.cxx and
> Testing/Code/Common/itkLandmarkBasedTransformInitializerTest.cxx
>
> The transform seems to have rotated the moving image in the wrong
> direction.
> Fixed image:
> http://crab-lab.zool.ohiou.edu/kevin/Slice_Landmark_Registration/cow_0.jpeg
> Moving image:
> http://crab-lab.zool.ohiou.edu/kevin/Slice_Landmark_Registration/cow_1.jpeg
> Difference of fixed and moving image:
> http://crab-lab.zool.ohiou.edu/kevin/Slice_Landmark_Registration/cow_diff_0_1.jpeg
> Difference of fixed and translated moving image:
> http://crab-lab.zool.ohiou.edu/kevin/Slice_Landmark_Registration/cow_diff_0_trans_1.jpeg
>
> Did I mess something up with the transform, centers of rotation and the
> like?
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
More information about the Insight-users
mailing list