[ITK-users] OrientImageFilter flipping VectorImage

mbcx9rb9 richard.brown at ucl.ac.uk
Mon Jun 11 10:17:59 EDT 2018


I have a multicomponent nifti image. It's a 3D volume and each voxel contains
3 values. I can then either save this to disk as a multicomponent nifti
image or as 3 single component images.

Using fslhd, I verify that all the headers of the single- and
multi-component files are identical, except for the obvious fact that one is
single component and the other is multi (e.g., dim0=3 vs. dim0=5). One thing
that is particularly important is that they both are in RAS, as shown in
this snippet:

          qform_xorient  Right-to-Left
          qform_yorient  Anterior-to-Posterior
          qform_zorient  Superior-to-Inferior

I read and reorient the images with a templated method:

          typedef itk::ImageFileReader<ITKImageType> ReaderType;
          ReaderType::Pointer reader = ReaderType::New();
          reader->SetFileName(filename);
          reader->Update();
          ITKImageType::Pointer itk_image_orig = reader->GetOutput();

          typename
itk::OrientImageFilter<ITKImageType,ITKImageType>::Pointer orienter =
                   
itk::OrientImageFilter<ITKImageType,ITKImageType>::New();
          orienter->UseImageDirectionOn();
          orienter->SetInput(itk_image_orig);
          orienter->SetDesiredCoordinateOrientation(
                   
itk::SpatialOrientation::ITK_COORDINATE_ORIENTATION_RAS);
          orienter->Update();
          ITKImageType::Pointer itk_image = orienter->GetOutput();
          std::cout << "\nflip axis:\n" << orienter->GetFlipAxes() << "\n";

Where the template is:
- itk::Image<float, 3> for the single component images 
or 
- itk::VectorImage<float, 3> for the multi component images

========
The problem
========
The images are given in RAS and the desired output is RAS, so no changes
should take place.
output for single component images: "flip axis: [0, 0, 0]"
output for multi component images: "flip axis: [0, 0, 1]"

This is causing problems as the origin moves accordingly for multicomponent
images. Is there any obvious reason as to why this is happening? Have I made
a mistake somewhere? Is this a bug?

Thanks



--
Sent from: http://itk-users.7.n7.nabble.com/


More information about the Insight-users mailing list