Hi SPMers.
I am trying to do the following:
(1) Downsample an image I to a lower resolution to give another image, lets call it 'D'
(2) Register a source image J to D to give a transformation T:J->D
(3) Apply T to image J to take it into the high resolution space of I.
I am using spm_reslice for step 1 to reslice I into the resolution of J which is low res.
Things go wrong when I try to apply the derived transformation obtained with the low res D image to transform between I and J - the images wont line up even though the low res D image and J can be lined up using T.
I am attempting to transform between I and J by applying the transformation matrix to the .mat file of I in the same way as one would do with D but it doesnt seem to work.
I can't work out why this should be, isnt 'T' a world to world coordinate transformation and so cant it be applied to resliced or upsampled versions of the same image?
Any ideas greatly appreciated.
Thanks
Anil
|