Hi,
also I would suggest to give mri_robust_register a try for the
registration. It is very accurate especially in cases with differences
in the images (jaw movement...). It can detect outlier voxel and remove
them from the registration.
Oh, and why do you use the T1.mgz? The norm.mgz is only brain and the
resulting registrations will be more accurate.
Cheers, Martin
On Tue, 2009-08-18 at 10:50 -0400, Bruce Fischl wrote:
> you can change the orientation without reslicing, but if the voxel
> dimensions of the 001.mgz are different from the orig.mgz then it will
> have to reslice. I don't see why you need to at all though - both contain
> ras2vox xforms that tell you how to initialize a transform.
>
> cheers
> Bruce
>
>
> On Tue, 18 Aug 2009, Renske de Boer wrote:
>
> > Hi Bruce,
> >
> > Thank you for your answer. What I forgot to mention is that I will not be
> > using FreeSurfer for the registration. We have our own registration method
> > that I prefer in this case.
> > I thought of using the -rl switch in mri_convert but as you mentioned this
> > will also resample the data. Do I understand correctly that it is not
> > possible to change the orientation in FreeSurfer without resampling?
> >
> > Renske
> >
> >
> > Bruce Fischl wrote:
> >> Hi Renske,
> >>
> >> you can use the -rl mri/orig/001.mgz switch in mri_convert ("reslice
> >> like"). Note that depending on the resolution etc. of the 001.mgz it may
> >> do
> >> an additional interpolation. You shouldn't need to do this though for
> >> registration as it should read this information from the header. Martin
> >> Reuter (ccd) has some very nice tools for this kind of thing.
> >>
> >> cheers,
> >> Bruce
> >>
> >>
> >> On Tue, 18 Aug 2009, Renske de Boer wrote:
> >>
> >>> Hi,
> >>>
> >>> I would like to register a FreeSurfer segmentation (or actually the
> >>> corresponding T1 image) to another scan of the same subject. In order to
> >>> perform this registration, the mri/T1.mgz file needs to have the same
> >>> orientation as the original T1 scan (mri/orig/001.mgz). Resampling is
> >>> not necessary since I want to prevent multiple interpolations and
> >>> resampling is already necessary in the registration/transformation.
> >>>
> >>> How can I convert mri/T1.mgz to the same orientation as the original T1
> >>> scan?
> >>>
> >>> Thanks!
> >>> Renske
> >>>
> >>>
> >
> >
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
