You're pretty much correct, but note that the information in the images
voxel-to-world transforms (from spm_get_space) are also used, so there may be
rotations etc that you need to take into account (if you use the SPM DICOM
utilities to import images that were not acquired in the usual transverse
orientation).
Best regards,
-John
> IN SPM2, when performing a rigid body six parameters registration (with
> spm_coreg), we can set up some options before the computation begins.
> When dealing with the starting parameters estimate:
> % params - starting estimates (6 elements)
> % default: [0 0 0 0 0 0]
> How could we set up "good" starting estimates? The first three are
> somehow related to the translations, and the last three to the
> rotations. But is there a physical interpretation of these parameters at
> this time? For example, if by some miracle we knew that our source image
> should by translated by, say, 20 mm on the X axis, and rotated by theta
> radians along X, how can we make these informations relevant before the
> computation? By making P = [20 0 0 theta 0 0 ] ?
|