> i want to segment a MRI image with custom atlas. i made the priors and
> smoothed them with 8m FWHM filter.
>
> i think i must not to smooth the atlas when using SPM5b, but if i have one
> image to create my one atlas, i think i must smooth it.
>
> but the results when i use this atlas are not good. !!!
>
> how i can correct it?
I would advocate the use of custom atlases much less for segmenting in SPM5,
and simply use the tpm data that comes with the distribution.
Best regards,
-John
|