Delete the y_* file if you have not done so yet. This provides starting
estimates, which will be based on the previous results.
Best regards,
-John
On Wednesday 07 January 2009 20:42, Kristopher Kosmatka wrote:
> Thanks very much John. That, in fact, was A problem. I used coreg reslice
> on the second image so that both images now have a negative x voxel size
> and are stored in RPI order. Unfortunately the same error in HDW still
> occurs. Any other ideas?
>
> Cheers,
> Kris
>
> On Wed, 7 Jan 2009 19:39:40 +0100, John Ashburner <[log in to unmask]>
wrote:
> >My best guess is that one of the images is stored left-handed, but the
> > other is stored right-handed. If you display an image, does one have a
> > negative x voxel size, but the other not?
> >
> >Best regards,
> >-John
> >
> >On Wednesday 07 January 2009 18:31, Kristopher Kosmatka wrote:
> >> I have been encountering an error with high dimensional warping, I hope
> >
> >someone may know a solution. I have two grey matter segment images that
> > have been previously rigidly coregistered. The images are unsigned INT
> > data type, 256x124x256 voxels. Nothing is unusual that I can see about
> > the images. I am running SPM5 revision 1782, in MATLAB 7.7, on x84_64
> > machine running Scientific Linux (An enterprise Red Hat clone). The
> > symptoms: All default settings were used in the HDW setup. The HDW
> > begins and prints the following
> >
> >to standard output:
> >> >> --------------------------
> >>
> >> Running "High-Dimensional Warping"
> >> Generating uniform affine transformation field
> >> Warping (iterations=8 regularisation=4)
> >> 193.63152
> >> nan
> >> XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX...
> >>
> >>
> >> This is followed by many X's ad infinitum. I've allowed this to
> >> continue running for 15-20 minutes and it just continues like this
> >> without ever finishing the job. Recompiling the mex files in my SPM
> >> distribution using the makefile in src failed to fix the problem. Has
> >> anyone encountered this problem, any ideas?
> >>
> >> Many thanks,
> >> Kris
> >>
> >>
> >> Kristopher J. Kosmatka
> >> Research Specialist
> >> Johnson Neuroimaging Lab
> >> University of Wisconsin - Madison
> >> Wm. S. Middleton Memorial VA Hospital - GRECC
|