[ITK-users] [ITK] Geodesic level set error

[Nicolas Gallego]

Hi Sindhoor,
Since you are changing the size of the images via the
RegionOfInterestFilter, then  you should check that the images provided to
your fast marching filter are on the same physical space.
this fragment of code taken form the
Examples/Segmentation/FastMarchingImageFilter.cxx
works because the image region size is that of the BufferedRegion of the
reader.
With the information you have provided that's what I can suggest to check,
if anybody has particular experience with this FastMarchingImageFilter that
should be taken into account it would be appreciated.

Nicolás Gallego-Ortiz
Université catholique de Louvain, Belgium


2014-06-17 14:07 GMT+02:00 Sindhoor Bhat <bhatsindhoor at ymail.com>:

> Hi Nicolas,
>
> Thanks for reply. This is what I did . Added below code to all filters &
> got origin & spacing of image
>
>                        OutputImageType:: Pointer image2=
> sigmoidFilter->GetOutput();
>                                image2->Print( std::cout ).
>
> But there is no problem with all the filters used. There is a problem
> when it comes to fast marching.
>  fastMarching->SetOutputSize(
> roi_filter->GetOutput()->GetBufferedRegion().GetSize() ); //took o/p from
> roi filter (after reading---ROI----gradient magnitude etc )
>
> * In itk manual I've one doubt (image attached). They say take buffer
> region from smoothing filter , but have taken o/p from reader. That is not
> clear.
>
> Can you pls tell what is wrong?
>
>
>   On Tuesday, 17 June 2014 2:26 PM, Nicolas Gallego <nicgallego at gmail.com>
> wrote:
>
>
> Hi Sindhoor,
>
> It's hard to tell. Could you provide a simple code that reproduces your
> error?
> A priori I would say you could print the image information on each
> connection of your pipeline, pay special attention to filters that may
> change it, as the ROI filter...
> Refer also to the concepts of regions: there are
> LargestPossibleRegion
> BufferedRegion
> RequestedRegion
>
> LargestPossibleRegion refers to the whole dataset. Refer to section 13.1
> of the software guide for more details.
> I hope that helps,
>
> Nicolás Gallego-Ortiz
> Université catholique de Louvain, Belgium
>
>
> 2014-06-17 10:36 GMT+02:00 Sindhoor Bhat <bhatsindhoor at ymail.com>:
>
> Hi nicgallego,
> Thanks for reply.
> I've read that section. Here is some detail on steps i followed. This may
> help you to give some hints.
>
> 1.read dicom series
> 2.define region of interest (ROI filter)
> 3. compute gradient & speed image.
> 4.Set seed using fast marching filter.
> 5.Give this inital level set image---- to geodesic segmentation filter
>
>   geodesicActiveContour->SetInput(  fastMarching->GetOutput() );
>   geodesicActiveContour->SetFeatureImage( sigmoidFilter->GetOutput() );
>
>
> fastMarching->SetOutputSize(roi_filter->GetOutput()->GetBufferedRegion().GetSize()
> ); //roi=region of interest
>
>  Even after reading from same dicom series, why is this error? it should
> have same origin / spacing right?
>  can you pls help me ?
>
>
>
>   On Tuesday, 17 June 2014 12:59 PM, Nicolas Gallego <nicgallego at gmail.com>
> wrote:
>
>
>  Good morning Sindhoor,
>
> In general itk pipelines process images in the physical space, so that
> pixelwise operations are between samples on the same physical position.
> Section 4.1.4 of the Software guide [1] and its correspondent example
> (Image4.cxx) may help you,
>
> [1] http://www.itk.org/ItkSoftwareGuide.pdf
>
> Nicolás Gallego-Ortiz
> Université catholique de Louvain, Belgium
>
>
> 2014-06-17 7:28 GMT+02:00 Sindhoor Bhat <bhatsindhoor at ymail.com>:
>
> Dear Matt,
>
> I've got error in Geodesic active contour level set segmentation :  saying
>  itk::ERROR: GeodesicActiveContourLevelSetImageFilter(0000000004AC7160):
> Inputs do not occupy the same physical space!
>
> I searched a lot on solving this error. but none have replied. Can you pls
> tell how to solve this error?
> (I'm trying to segment 3D dicom images)
>
> thank you.
>
> _____________________________________
> 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://www.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-users
>
> _______________________________________________
> Community mailing list
> Community at itk.org
> http://public.kitware.com/mailman/listinfo/community
>
>
>
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/insight-users/attachments/20140617/a71a9d3e/attachment.html>