[vtkusers] Displaying Oblique Orientation Image Correctly?

[Constantine Zakkaroff]

Hello ALL.

Can someone please help me out with a clarification?

Say, a volume with double-oblique orientation (short-axis cardiac stack) 
from DICOM is loaded. Then itk::ImageToVTKImageFilter<SomeImageType> is 
used to obtain vtkImageData. So far so good.

Now, since vtkImageData has no knowledge of the original oblique 
orientation of the image as it exists in itk::OrientedImage<TPix, TDim>, 
when displaying the resulting vtkImageData in vtkImageViewer2 none of 
the voxels appear in their correct physical location apart from the 
voxel at vtkImageData origin. This is (obviously) because advancing 
along the grid of our vtkImageData has a different/incorrect meaning 
without the original double-oblique orientation; the grid of 
vtkImageData does not correspond to the original physical space of the 
image.

I realise one way to display an image like that is to use vtkImageActor 
with an appropriate transform set to it, but this is not suitable when 
visualising registration outcomes with vtkImageCheckerboard.

What is the process of displaying such images correctly? Is this is the 
case where one would use vtkGridTransform? Or is there something else 
I'm missing?

Many thanks,
Constantine