[Insight-developers] Image::DirectionType should be Nx3 ratherthan NxN?

Kris Thielemans kris.thielemans at csc.mrc.ac.uk
Tue Sep 18 12:18:48 EDT 2007


Hi

> 
> On Mon, Sep 17, 2007 at 08:54:28AM -0400, Luis Ibanez wrote:
> >
> > Hi Steve,
> >
> > We are following up on the bug report:
> >
> >   "5573: Image::DirectionType should be Nx3 rather than NxN"
> >   http://public.kitware.com/Bug/view.php?id=5573
> >
> > We are tempted to mark this as a "will not fix" item.
> >
> > The reason being that when you read a 2D DICOM slice that 
> actually has 
> > 3D orientation information, such image should be read into an ITK 
> > image type instantiated for 3D, not an image type 
> instantiated for 2D.

this is a very difficult decision I think. I tend to agree with Steve.
However, I don't think this can be fixed for everyone. Say you have
microscope images, or projection images (such as x-ray). for those, the 3rd
component of the origin doesn't make sense anyway. 

I guess one could solve this by having the Image::DirectionType dimensions
depending on another template parameter, defaulting to 3 (or N).

Kris Thielemans
Hammersmith Imanet Ltd, part of GE Healthcare 



More information about the Insight-developers mailing list