the EMS-package from Leuven made an attempt to implement an automated lesion detection by a MRF-based algorithm. You may want to check how it performs on your data.
Von: Philipp Sämann [mailto:[log in to unmask]]
Gesendet: Do 30.09.2004 19:12
An: [log in to unmask]
Betreff: [FSL] Automated lesion segmentation with FAST
Hello FSL users,
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!
Max-Planck-Institute of Psychiatry
NMR Study Group