Hi,
There are many other options in BET that you could try, but I suspect the best things to try are:
(a) crop your original images to remove excess neck (this can be automated with the tool robustfov)
(b) try fsl_anat instead (you can use --noseg and --nosubcortseg)
The second option uses a completely different method, based on non-linear registration, but is a lot slower so is not a good option if speed is a major priority.
All the best,
Mark
On 5 Jan 2013, at 16:41, Nucleos P <[log in to unmask]> wrote:
> Dear Experts,
>
> I have run BET on my datum and I used the -R and -m options. The sample size is around 400. Using the -R works well for most of the subjects but with others, a lot of the brain is being stripped away.
>
>
> Any suggestions on how to refine the use of the BET tool to get the best results?
>
> Thank you so much,
> Catherine
|