Hi,
You may want to use pencil tool to circle out the outliers, then use
fill tool to fill in the circles you've draw. This would give you a
mask with 1 for outliers. Then use fslmaths yourmask.nii.gz -sub 1
-mul -1 brainmask.nii.gz to generate a maks with 0 for outliers.
Finally, use fslmaths your_brain.nii.gz -mul brainmask.nii.gz
your_new_brain.nii.gz to generate the betted brain without the
outliers.
However, depends on the data, it may be a huge time & labor consuming
task, in my data, I need to spend 2-3 hrs to generate such a much:( My
friends suggest me to use recon-all from freesurfer do this job
automatically by using the autorecon1 procedure. I found it maybe a
nice way out.
On Tue, Jul 20, 2010 at 6:47 PM, Rob H <[log in to unmask]> wrote:
> Hi
>
> I am having some problems with brain extraction. We have run BET many times using various parameters and can't get it to successfully remove the skull from a T1 Flair image. Whatever we do we always have some fragments of skull left.
>
> Somone suggested creating a mask in fslview and using the eraser to effectively 'rub out' the sections of skull that remain from the BET extracted image by setting their intensity to 0. When I load the flair image in fslview all the voxels have intensity values. However when I select 'create mask' all the intensity values are set to 0. Furthermore I can't seem to get the Eraser tool to do anything. I can highlight areas with it, but it doesn't change the intensity of them, or remove them from the image.
>
> How are you supposed to use the eraser tool?
> What is the best way of removing remaining fragments of the skull after using BET?
>
>
> Thanks
> Rob
>
--
Best Regards
Xue, Feng Ph.D.
Major in Developmental Cognitive Neuroscience
National Key Laboratory of Cognitive Neuroscience and Learning
Beijing Normal University
Beijing, China. 100875
Tel: +86-13810154455
web: http://psychbrain.bnu.edu.cn
==============================================
Welcome to MuDuo JinSheng BBS @ Beijing Normal University
telnet://bbs.mdjs.org
http://bbs.mdjs.org
|