That appears to work perfectly, thanks!
I’m confused, though, as to me the direction of the operations appears to be reversed (the reg.dat file was obtained by moving the hires MRI to match the low-res surface). How do I know which one to use?
Date:
Wed, 14 Jun 2017 16:56:17 +0200
From:
Douglas Greve <greve@nmr.mgh.harvard.edu>
Subject:
Re: [Freesurfer] Applying an affine transformation to a
surface
To: freesurfer@nmr.mgh.harvard.edu
Message-ID:
<ba337251-ab46-8ef2-a1f3-25dbbd353776@nmr.mgh.harvard.edu>
Content-Type:
text/plain; charset="windows-1252"
Can
you try it with --reg instead of --reg-inv ?
On
6/13/17 8:55 PM, Daniel Gallichan wrote:
Hello
In a very similar situation to a post I found in the Freesurfer
archive, I have low-resolution (3T) and high-resolution scans (7T) of
the same subject - and would like to put the surfaces from the
low-resolution scan into the space of the hi-res data. The
instructions from the previous thread say to check example 4 when
running ?mri_surf2surf --help? - which I tried, but I cannot get the
result to overlap in freeview.
I have a registration file which looks very good from running
bbregister on the high-res data into the low-res space - but I cannot
perform the registration the other way around as I?m still having
trouble getting the hi-res data to process properly through recon-all
(it?s acquired at 350um, and I?m currently running it downsampled to
500um, but that?s still a work-in-progress?).
So this is the command-line that I tried to apply the transformation
from the registration to a surface:
mri_surf2surf --reg-inv hires_to_lowres.dat
${SUBJECTS_DIR}/subj_hires/mri/T1.mgz --hemi lh --sval-xyz white
--tval-xyz ${SUBJECTS_DIR}/subj_hires/mri/T1.mgz --tval lh.white.hires
--s subj_lowres
If I then try to view the result:
freeview -v ${SUBJECTS_DIR}/subj_hires/mri/T1.mgz -f
${SUBJECTS_DIR}/subj_lowres/surf/lh.white.hires
it is clear that the two are far from being aligned?
I?m afraid I may be misunderstanding how to control the coordinate
systems in the mri_surf2surf command - especially as the source and
targets are different ?subjects? (the same actual subject, though?). I
can?t used the ?--trgsubject? option because recon-all has not
completed properly on the hires data, and there is no sphere.reg file
(although I have to confess, I haven?t understood why it would need it
if I am specifying my own registration file?) . Any suggestions very
gratefully received.
Thanks
Daniel
*Dr Daniel Gallichan*
*Lecturer*
CUBRIC
School of Engineering
Cardiff University
Maindy Road
Cardiff
CF24 4HQ
Wales, United Kingdom
Tel: +44 (0) 29 208 70045
Email: gallichand@cardiff.ac.uk <mailto:gallichand@cardiff.ac.uk>
*Dr Daniel Gallichan*
*Ddarlithydd*
CUBRIC
Ysgol Beirianneg
Prifysgol Caerdydd
Heol Maindy
Caerdydd
CF24 4HQ
Cymru, y Deyrnas Unedig
Ff?n: +44 (0) 29 208 70045
E-bost: gallichand@caerdydd.ac.uk <mailto:gallichand@caerdydd.ac.uk>
Personal website:
http://www.cardiff.ac.uk/people/view/507850-gallichan-daniel
CUBRIC - *C*ardiff *U*niversity *B*rain *R*esearch *I*maging
*C*entre - http://sites.cardiff.ac.uk/cubric/
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
--------------
next part --------------
An
HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/freesurfer/attachments/20170614/bb4e212f/attachment-0001.html