[Insight-developers] bug in BSplineDeformableTransform?

[Luis Ibanez]

Stehan, Tomas,

What we found is that the problem was related to the computation
of the start index used as the first node of support in the
BSpline grid.

A Fix has been committed to the CVS repository:
http://www.itk.org/cgi-bin/viewcvs.cgi/Code/Common/itkBSplineInterpolationWeightFunction.txx?root=Insight&r1=1.13&r2=1.14&sortby=date 


Tests were added for 1D resampling using Spline orders 1,2 and 3:
http://www.itk.org/cgi-bin/viewcvs.cgi/Testing/Code/Common/itkBSplineDeformableTransformTest2.cxx?root=Insight&sortby=date&view=log 


A test of kernel symmetry has been added:
http://www.itk.org/cgi-bin/viewcvs.cgi/Testing/Code/Common/itkBSplineInterpolationWeightFunctionTest.cxx?root=Insight&r1=1.4&r2=1.6&sortby=date 


The bug 4021 has been marked as "resolved":
http://public.kitware.com/Bug/view.php?id=4021

Please give it a try and let us know if this works for you,
in which case we will proceed to close the bug.


    Thanks


      Luis


--------------------
Luis Ibanez wrote:
> Stefan, Tomas,
> 
> Thanks for pointing this out and for proposing
> the fix.
> 
> We will take care of it before cutting ITK 3.6.
> 
> 
>   Thanks
> 
> 
>       Luis
> 
> 
> ---------------------
> Tomáš Kazmar wrote:
> 
>> Hi Stefan,
>>
>>  thanks for sending the link, I attached some more info to the bug 
>> report and
>> hopefully the right way to patch it.
>>
>> Regards,
>> Tomas
>>
>>
>> # Hi Tomas,
>> # # This seems to be the same problem as bug 4021, which i reported a 
>> while ago.
>> # # http://www.itk.org/Bug/view.php?id=4021
>> # # Apparently it hasn't been fixed yet. :)
>> # # Kind regards,
>> # Stefan.
>> # # At 23:12 18-3-2008, =?iso-8859-2?Q?Tom=E1=B9=20Kazmar?= wrote:
>> # >Hi all,
>> # >
>> # >   recently i tried to use second order b-splines to warp images and
>> # >the results contained sharp steps (~were not smooth). I am not sure
>> # >whether I do something wrong. Has anybody used it successfully?
>> # >
>> # >It seems to me that there is an error in
>> # >BSplineInterpolatingWeightFunction in combination with
>> # >BSplineKernelFunction. The first uses kernel in the range -1..2 and
>> # >the second defines the support as -1.5..1.5.
>> # >
>> # >Should I report it as a bug? There is a simple patch but once more
>> # >I am not sure whether I use it the right way.
>> # >
>> # >Here is the relevant part of code:
>> # >
>> # >   const unsigned int SplineOrder = 2;
>> # >   typedef itk::BSplineDeformableTransform<float, Dimension, 
>> SplineOrder> # > TransformType;
>> # >   TransformType::Pointer bsplineTransform = TransformType::New();
>> # >   typedef TransformType::RegionType RegionType;
>> # >   RegionType bsplineRegion;
>> # >   RegionType::SizeType size;
>> # >   const unsigned int nodesOutsideImage = SplineOrder;
>> # >   const unsigned int nodesInsideImage = 4;
>> # >   const unsigned int numberOfGridNodes = nodesInsideImage + # > 
>> nodesOutsideImage;
>> # >   size.Fill( numberOfGridNodes );
>> # >   bsplineRegion.SetSize( size );
>> # >   typedef TransformType::SpacingType SpacingType;
>> # >   SpacingType spacing;
>> # >   spacing[0] = 1.0 * floor( (width-1) / 
>> (double)(nodesInsideImage-1) );
>> # >   spacing[1] = 1.0 * floor( (height-1) / 
>> (double)(nodesInsideImage-1) );
>> # >   typedef TransformType::OriginType OriginType;
>> # >   OriginType origin;
>> # >   origin[0] = 0.0 - spacing[0];
>> # >   origin[1] = 0.0 - spacing[1];
>> # >   bsplineTransform->SetGridSpacing( spacing );
>> # >   bsplineTransform->SetGridOrigin( origin );
>> # >   bsplineTransform->SetGridRegion( bsplineRegion );
>> # >
>> # >
>> # >Regards,
>> # >Tomas
>> # >_______________________________________________
>> # >Insight-developers mailing list
>> # >Insight-developers at itk.org
>> # >http://www.itk.org/mailman/listinfo/insight-developers
>> # # # # _______________________________________________
>> Insight-developers mailing list
>> Insight-developers at itk.org
>> http://www.itk.org/mailman/listinfo/insight-developers
>>
>