[ITK-dev] BSplineTransform TransformPoint() crash
Bradley Lowekamp
blowekamp at mail.nih.gov
Wed Apr 20 10:18:17 EDT 2016
Hello,
Thank you for reporting this issue and providing a clear example and proposed solution.
In looking at the problematic function[1], I a quite surprised to see that the the signature of the method:
bool BSplineTransform::InsideValidRegion( ContinuousIndexType & index ) const
Specifically, the argument is not constant and it is being modified by this problematic case.
I am suspicion of the need for this if statement in the method. It appears there may be a deeper issue with how the b-spline is defined over the upper range of the last index. The transform should be valid from (minIdx-0.5, maxIdx+0.5). I am going to also note that this case is not in itkBSPlineDeformableTransform.
Brad
[1] https://github.com/InsightSoftwareConsortium/ITK/blob/master/Modules/Core/Transform/include/itkBSplineTransform.hxx#L465-L495
> On Apr 20, 2016, at 7:07 AM, Denis Shamonin <dshamoni at gmail.com> wrote:
>
> Dear Developers,
>
> We have found the rare but reproducible crash during the resampling in combination with BSplineTransform. Could you please review this issue on JIRA? More details here:
> https://issues.itk.org/jira/browse/ITK-3425 <https://issues.itk.org/jira/browse/ITK-3425>
> Thanks,
>
> Denis P. Shamonin, MSc
> Division of Image Processing (LKEB)
> Department of Radiology
> Leiden University Medical Center
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://kitware.com/products/protraining.php
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://public.kitware.com/mailman/listinfo/insight-developers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/insight-developers/attachments/20160420/7acac0b4/attachment.html>
More information about the Insight-developers
mailing list