Hi,
I would try a cost function that does not rely on binning (normalised
correlation if they are the same modality) or reduce the number of
bins for the other cost functions (correlation ratio or mutual informations).
All the best,
Mark
On 29 Jun 2011, at 22:43, Blake Evans Dewey wrote:
> Ok. So my images are very small (64x64). How should I adjust my settings to correct for that?
>
> ----------------------------
> Blake Dewey
> Vanderbilt University, 2013
> Physics and Mathematics
>
>
> On Wed, Jun 29, 2011 at 4:41 PM, Mark Jenkinson <[log in to unmask]> wrote:
> Hi,
>
> You cannot impose a limit on the transformation parameter values.
> If it is finding very large values then there is no point in putting a
> limit on them because it will just choose this limit, which is arbitrary,
> as it must believe that the cost function gets better and better the
> further away it tries.
>
> The question really is why is it doing this, and it might be worth
> trying other cost functions or parameters (like the number of bins)
> since a small 2D image would have a very different number of voxels
> from a typical 3D image, and that might be causing the problem.
>
> If you are still having trouble then feel free to upload the images
> to our uploading website and we can have a look at them.
>
> All the best,
> Mark
>
>
> On 29 Jun 2011, at 19:45, Blake Evans Dewey wrote:
>
> > Thanks. Directory was misspelled. Also is there any way to restrict the translations that are done. My images are close to alignment (with signal differences) but FLIRT tries to transform them almost out of the FOV.
> >
> > ----------------------------
> > Blake Dewey
> > Vanderbilt University
> > Physics and Mathematics
> >
> >
> > On Wed, Jun 29, 2011 at 12:10 PM, Blake Dewey <[log in to unmask]> wrote:
> > I have written two 2D images from MATLAB using the NIFTI toolbox. However when I try to run FLIRT with the -2D option it reports that it cannot read the header file of my .nii file. I have also tried creating the image using MIPAV.
> >
> > flirt -in input -ref ref -out temp -omat temp.mat -2D
> > ** ERROR (nifti_image_read): failed to find header file for 'ref'
> > ** ERROR: nifti_image_open(ref): bad header info
> > Error: failed to open file ref
> > ERROR: Could not open image ref
> > Image Exception : #22 :: Failed to read volume ref
> > terminate called after throwing an instance of 'RBD_COMMON::BaseException'
> > Abort trap
> >
> > Any suggestions?
> >
> > Blake Dewey
> > Vanderbilt University Institute of Imaging Science
> >
> >
>
>
|