Ken,
I encountered a similar problem. In my case, I have done bias correction of T1
image with N3. When I try to segment the image corrected with N3, it gives error
message, "Warning: Matrix is singular, close to singular or badly scaled.
Results may be inaccurate. RCOND = NaN." and processing terminated.
Checking the images, I noticed one thing. The scaling factors of bias corrected
images with N3 are very low. I thought this could be a problem and I found some
notes regarding scaling factor in SPM in the following links.
http://www.matthijs-vink.com/minc.html
http://imaging.mrc-cbu.cam.ac.uk/imaging/FormatAnalyze
In my case, I set the scaling factor of those images to 1 with MRIcro and
re-tried the segmentation with SPM5, which resulted in complete success. So it
might be good to check the scaling factor of your bias-corrected images just in
case.
Hope this helps,
Kiyotaka
On Wed, 20 Jun 2007 15:11:08 +0100
Kenneth Rando <[log in to unmask]> wrote:
> Hello all,
>
> I have T1 images that have already been bias corrected using another
> method. When I try to segment in SPM5 using "no correction" I receive
> the "matrix singular to working precision or badly scaled" error message
> and processing terminates. Is this a bug in the code, or do I need to
> change other parameters? As a test I have segmented the same images using
> 150mm and the default regularization setting. The processing completes
> sucessfully. I am concerned, however, of the effect of running a bias
> correction procedure on data already bias corrected using a different
> method.
>
> Thanks.
>
>
> Ken Rando
----------------------------
Kiyotaka NEMOTO, M.D.
Mitsukaido Kosei Hospital
3721-2 Uchimoriya Joso Ibaraki
300-0043 Japan
E-mail: [log in to unmask]
----------------------------
|