Hi,
Before the release we put some effort into speeding up the
inverse warp procedure and it seemed to only take 15 minutes
or so on our machines. Can you send me some information
about the images you are using? Particularly the size in
voxels and mm. Hopefully we can figure out why it is taking
so long and do something to avoid it.
All the best,
Mark
On 3 Sep 2008, at 15:42, Vishwadeep Ahluwalia wrote:
> Hi,
> I ran featquery on a 1st level FEAT which used nonlinear
> registration. the
> nonlinear inverse warping took overnight to complete. Has anyone had
> the
> same issue? is this plausible? and is there a workaround to save
> time? i
> could still use the linear registration ('reg-' directory) to speed up
> things but i'm guessing for subcortical rois, using non-linear inverse
> registration would be more accurate than linear...wouldnt it ? any
> help is
> appreciated.
> cheers,
> Vish
>
|