If you define a bug as an unintended behaviour or result of a computer
program, I don't see how holes in a mask resulting from a brain
extraction tool can be called anything else.
I see this a lot with pre-extracted brains. I work around it by
applying bet twice (or more often) with slightly varying -f, and
combining the results with 'fslmaths brain1 -max brain2 finalbrain'.
When I need results to be robust, I apply a label propagation
procedure instead of BET, akin to Leung et al. ("Brain MAPS",
Neuroimage 2011). This is computationally expensive, but one gets
fewer unwelcome surprises.
Regards
Rolf
On 17 January 2012 09:59, Stephen Smith <[log in to unmask]> wrote:
> Hi - unlikely to be a bug; the BET optimisation procedure involves
> deforming a mesh surface, and this process is a complex optimisation
> (including detecting and fixing self-intersections), so it is quite possible
> for some settings to give different results from others - although you're
> right that in general the results vary smoothly as a function of the input
> parameters.
>
> Cheers.
>
>
>
> On 16 Jan 2012, at 23:19, Mishkin Derakhshan wrote:
>
> Hi,
> For some reason, on a subset of our images, when we use -f 0.35 we end
> up with a hole in the cerebellum.
>
> screenshot: http://imgur.com/tuivH
>
> If we use -f 0.349 or -f 0.351 then there is no such hole.
>
> The input image is already an extracted brain. ie. we are doing bet
> twice as it cleans up the images a bit better.
> I know I am using an older version of bet, and I've already found a
> solution of just not using -f 0.35 exactly but to add or subtract
> 0.001, but this seemed like a bit of a bug.
>
> Any idea of what is going on? Was this fixed in a later version?
>
> thanks,
> mishkin
>
>
>
> ---------------------------------------------------------------------------
> Stephen M. Smith, Professor of Biomedical Engineering
> Associate Director, Oxford University FMRIB Centre
>
> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
> +44 (0) 1865 222726 (fax 222717)
> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
> ---------------------------------------------------------------------------
>
>
>
--
Rolf A Heckemann, MD PhD
Médecin chercheur
Fondation Neurodis
CERMEP - Imagerie du Vivant
Hôpital Neurologique Pierre Wertheimer
59 Boulevard Pinel
69003 Lyon
France
|