Hi Xavier.
It is quite possible that even though you have a t1-weighted scan it still
has quite different contrasts and intensities to the average image - so
least squares may not be a good cost function (ie may not allow enough
flexibility in the intensity transformation from your image to MNI).
In general correlation ratio works best (as you found) - why did you want
to use least squares?
Thanks, Steve
On Tue, 21 May 2002, Xavier Chitnis wrote:
> Hi,
>
>
>
> I have been using Flirt to map T1 MRI scans to the avg 152 brain
> template. When I use the correlation ratio or normalised correlation as
> the cost function the registration seems to work fine. However, when I
> tried to use least squares (which is what I would ideally like to use),
> the images end up being turned into a different orientation when I view
> them (in Mricro). I chose the option that images were not aligned but
> were the same orientation.
>
>
>
> Any help would be greatly appreciated,
>
>
>
> Xavier
>
>
>
>
>
> Xavier Chitnis MSc.
>
> Neuroimaging Research
>
> Department of Neurology
>
> Institute of Psychiatry
>
> De Crespigny Park
>
> London, SE5 8AF
>
>
>
> 020 7848 0403
>
>
>
>
Stephen M. Smith
Head of Image Analysis, FMRIB
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|