Hi Tine,
the problems you see seem to be due to bad initial image alignment,
which again, in my experience, is inhomogeneity unless proven otherwise.
Hence, I would try allowing more bias correction by fiddling with the
respective parameters in unified segmentation (mainly bias
regulirazation and bias FWHM) to see if that helps. Sometimes a two-pass
procedure is a last resort, so that you "segment" the image only to get
a bias-corrected version and then segment that one again.
Hope this helps,
Marko
Tine Hansen wrote:
> Hi,
>
> I just started using SPM and have some questions regarding the
> segmentation of structural images.
>
> I did segmentation of my structural image using default settings and got
> the results as seen in the attached
> file 'structural_segmented'. The top right image is the original
> structural image. The bottom image is the normalised image
> (using seg_sn.mat and the bias corrected image), which of cause is bad
> due to bad segmentation.
>
> I manually reoriented the structural image using "Display" where I
> changed the values "right, forward, up,
> pitch, roll and yaw". I then used "Reorient images" and chose my
> structural image. I did the segmentation on the reoriented image
> and got results as shown in the attached file
> 'reoriented_structural_segmented'. The segmented images are still bad.
>
> Why do a get such a bad segmentation and what can I do to get a better
> segmentation?
>
>
> Thanks, Tine
--
=====================================================================
Marko Wilke (Dr.med./M.D.)
[log in to unmask]
Universitäts-Kinderklinik University Children's Hospital
Abt. III (Neuropädiatrie) Dept. III (Pediatric neurology)
Hoppe-Seyler-Str. 1, D - 72076 Tübingen
Tel.: (+49) 07071 29-83416 Fax: (+49) 07071 29-5473
=====================================================================
|