Dear Sean,
The naming of the FAST outputs depends on the intensity values - they are ordered by mean intensity.
If you want to find a consistent partition across all then I suggest you look at the spatial correlations between your partitions in different subjects, or correlations to a standard template. That should let you pick something consistent within your cohort.
All the best,
Mark
On 7 Dec 2011, at 12:47, Sean F Walsh wrote:
> Hi Experts,
>
> I have a group of stroke patients with large lesions. I want to perform 4-part FAST segmentation on the data (as I don't want the lesion to be included in the white or grey matter sections). I have painted lesion masks for each subject. I then multiplied the lesion masks by 100 and multiplied (lesionx100)x(original_T1) to output a T1 for each subject with very high intensity values in the lesion.
>
> FAST seems to work well on this, but there is something rather annoying about it. I have scripted the process (as part of a larger script) but sometimes the grey matter is outputted as fast_seg1 and sometimes as fast_seg2. Is there any way to ensure consistency of FAST output naming when dealing with 4 partitions?
>
> Many thanks in advance,
>
> Seán
>
|