[Insight-users] Bug in MatrixOffsetTransformBase<..., ..., ...>::GetJacobian

[Luis Ibanez]

Hi Elena,


Regarding the issue that you reported: Bug: #4401

http://public.kitware.com/Bug/bug.php?op=show&bugid=4401&pos=0

A fix for the MatrixOffsetTransformBase has been committed to
the MAIN branch of the CVS repository:
http://www.itk.org/cgi-bin/viewcvs.cgi/Code/Common/itkMatrixOffsetTransformBase.txx?root=Insight&sortby=date&r2=1.13&r1=1.12 


Note that the regression testing image for ImageRegistration9Test
was also updated, due to minor differences in the mapping of pixels
in the border of the image.


Please give it a try and let us know if you find any problems.


   Thanks


     Luis


---------------------------
Elena Pavlovskaia wrote:
> Hi,
> 
> I think that
> 
> MatrixOffsetTransformBase<TScalarType, NInputDimensions, NOutputDimensions>
> ::GetJacobian( const InputPointType & p ) const
> 
> has a bug. It does not use the transform center though it should use it.
> 
> Here are the indications and the fix:
> 
> I was running image registration with MatrixOffsetTransformBase using very
> similar images. With every next iteration the registration became worse. The
> same images were registered fine when I used Similarity3Dtransform.
> 
> I modified the code in MatrixOffsetTransformBase<...>::GetJacobian so that
> to use center (similar to Similarity3Dtransform):
> 
> In the beginning of the function I added
> 
>   const InputVectorType pp = p - this->GetCenter();
> 
> and then used pp instead of p in the loop.
> 
> After this change my image registration works fine.
> 
> Thanks,
> 
> Elena
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>