[MINC-users] MINC convention for oblique imaging

[Peter Neelin]

I'm a little behind on my BIC e-mail, but I hope that I am not too
late to comment:

On 7/23/08, Andrew Janke <a.janke at gmail.com> wrote:
> > By MINC definition, xyz dimensions are defined in patient coordinates
>
>
> Um maybe? :)  I always was of the opinion that the world co-ordinates
>  referred to the scanner. And thus 0,0,0 for a MRI machine is at the
>  centre of the gradients and 0,0,0 on some PET machines is at the
>  patients feet.

Leila is right. Here's a quote from the 1.4 programmer's guide:

The MINC standard defines how spatial coordinates should be oriented
relative to patients. Files are free to have data stored in the
desired direction, but positive world coordinates are given a definite
meaning in the medical imaging context. The standard is that the
positive x axis points from the patient's left to right, the positive
y axis points from posterior to anterior and the positive z axis
points from inferior to superior.


>  > Therefore, in an oblique dataset, do MINC xyz dimensions refer to
>  > patient coordinates or MRI voxel coordinates?

The xyz dimensions are obviously along the sampling axes, but the
labels are chosen by finding the patient axis that is closest to the
sampling axis. You can, of course come up with pathological cases that
are ambiguous (especially with non-orthogonal axes), but then you just
have to break the tie. You should only use a label once (don't use two
x axes and no y axis).

> The answer to this is to just think of the MINC world space as a
>  sampling of the scanner, if a patient head happens to be in there then
>  good and well. :)  Of course this does beg the question of why there
>  are -coronal type options in MINC but my answer to this would be "ease
>  of use for 90% of users"

Actually, it's software developer laziness (me included) that leads to
using scanner coordinates. One should take into account whether the
patient is prone or supine, etc., but frequently software just uses
the scanner axes 'cause it's easier and people usually lie on their
backs (brains without bodies are another matter). Note that if you are
converting from DICOM, then you can just use Image Position Patient
and Image Orientation Patient since these are (fairly evidently)
patient-based, or should be if properly implemented and the patient
orientation is set correctly during the scan.

Peter
-- 
Peter Neelin
(peter.neelin at gmail.com)