Hi,
See attached for some modifications to the sienax script that allows
for you to pass previously extracted brain/skull images using the -P
option (-P brain-image skull-image). I have a similar modification for
siena, if needed.
Generating a skull image from an already existing brain image isn't
possible. A co-worker and I played around with this for a couple hours
some years ago. We made some modifications to the bet source code to
allow for this but we weren't quite happy with the results and didn't
pursue it further. In general, unless the skull image is *drastically*
wrong, I don't think you'll find that the vscaling will be so
different anyway if you try different skull images for a given brain
extraction. Remember, the skull image is an estimate heuristically
derived due to the difficulty in distinguishing between skull/CSF
using just a T1-weighte image.
By the way, Mark Jenkinson replied to a similar question earlier and
said they are working on improved brain extraction.
On Fri, Jan 24, 2014 at 10:31 AM, Gabor Perlaki
<[log in to unmask]> wrote:
> Dear all,
>
> Is it possible to manually refine the Bet in Sienax? My problem is that the Bet also estimates the skull image and if the Bet is not good then the skull image will also be probably not perfect. Is it possible to create a skull image initialized by the manually refined brain image?
>
> Thanks,
> Gabor
--
Niels Bergsland
Integration Director
Buffalo Neuroimaging Analysis Center
100 High St. Buffalo NY 14203
[log in to unmask]
(716) 859-7677
|