Hi - the problem with using a generic segmentation method for lesions is
that most methods (like FAST) look for clusters in the histogram to
separate as a separate class (assuming it knows the right number to look
for) and lesions can go from large to very small in total image voxel
count, so this isn't a reliable method. We've been looking into a script
that combines separate segmentations of different modalities but this
isn't terribly reliable yet.
Cheers, Steve.
On Thu, 30 Sep 2004, Philipp S[ISO-8859-1] ämann wrote:
> Hello FSL users,
> hello Steve,
>
> is there a good combination of modalities/options with FAST or multichannel-
> FAST that could automatically segmentate hyperintense lesions as typically
> seen with multiple sclerosis or other white matter disease?
>
> The data pool of interest offers T1 (high res), T2, PD and FLAIR (axial,
> gapless) images. So far, some trial-and-error attempts did not produce
> lesion masks, but lesions were rather added to grey matter or CSF.
>
> Thanks a lot!
> Philipp
>
> Max-Planck-Institute of Psychiatry
> NMR Study Group
> Kraepelinstr. 2-10
> 80807 Munich
>
Stephen M. Smith DPhil
Associate Director, FMRIB and Analysis Research Coordinator
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|