In order for this to work well, the distribution of intensities in the
templates must be the same as (or similar to) the distribution of intensities
in your images. For example, if the CSF is dark in your images, then it
should also be dark in the template. This is because the algorithm tries to
work by minimising the residual sum of squares difference.
The segmentation button of SPM5 does not have the same constraints. All it
needs is for the images to have GM, WM and CSF, each with different intensity
distributions. This is one of the reasons why it generally outperforms the
results obtained by the Normalise button.
Best regards,
-John
On Tuesday 02 September 2008 19:07, Cibu wrote:
> Hello
>
> I am trying to normalize single subject FA volumes on to the T2 template. I
> have done this before with spm 2 and the normalization worked beautifully.
>
> Now, I am getting badly warped normalization output and would greatly
> appreciate your help with fixing this.
>
> I presumed that the warping may be because of the difference in orientation
> of the FA volume and flipped it along x and y planes to correspond with the
> orientation of the T2 template, which I am told is in Neurological
> convention. But this didn't help either.
>
> Attached are some pdf files that show:
>
> 1. the original FA volume (we used a vitamin e capsule on the right temple
> area to avoid r/l confusions)
>
> 2. Flipped FA volume (note the position of the capsule shows that right is
> right)
>
> 3. Normalized output.
>
>
>
> Please note that I am still using SPM 2.
>
>
>
> Looking forward to hearing from you
>
>
>
> Cibu Thomas
>
> Post Doctoral Research Fellow
> Athinoula A. Martinos Center for Biomedical Imaging,
>
> Department of Radiology, Harvard Medical School,
>
> Massachusetts General Hospital,
>
> 149, 13th Street, Charlestown, MA 02129
>
> Tel: 617-726-9034
|