Dear Hauke,
When you talk about the "origin" do you mean the origin values stored
in the header file, or just the centre of the image?
I will assume here that you want the centre of the volumes to "coincide".
Given that your FOV in x and y are the same for both it means you only
need to make the z FOV bigger. In this case you need to shift the
z axis by half of the extra FOV. As you'll want to maintain your EPI
slice thickness I suggest making a dummy reference volume with
64x64x37 voxels of 3.75x3.75x5 mm size. This gives you the same
FOV as the T1 (well, 185mm instead of 186mm). The extra FOV in z is
(185-125) = 60mm. So just add 30mm to the z translation.
That is, add 30 to the third row, fourth column of the
flirt matrix. So assuming you saved the matrix with the -omat
option originally (if not, use convert_xfm to get this from a MEDx form)
then the file epi2anat.xfm should be a 4 by 4 ascii matrix.
Adding 30 to the (3,4) element and using a reference volume with
64x64x37 voxels of 3.75x3.75x5 mm size should give you what you want.
Obviously adding more to this (3,4) element will shift more, so if the
transformation is still cropping, then adjust this value until it gives
you what you want. However, if it isn't that close right away then the
transformation matrix might not be good.
Cheers,
Mark
P.S. If it really is the "origin" in the header file that is the problem,
then you need to create a dummy volume with the correct origin values
initially. See avwcreatehd.
Hauke Heekeren wrote:
>
> Hi Mark,
> I tried to make the appropriate changes of the translation parameters
> and get it working somehow.
> However I have not yet figured out the proper way of maintaining the
> origin.
>
> Here are some more details:
> In this specific case the T1 are 256x256, 0.9375x0.9375, 124 slices,
> 1.5mm (ie FOV 240x240x186).
> The EPIs are 64x64, 3.75x3.75 mm, 25 slices, 5 mm (ie FOV 240x240x125).
>
> Best,
>
> Hauke
|