Hi Seth,
Happy New Year to you too...
after having looked at one of your problematic dataset
(#930467), it appears that FAST is indeed struggling
with the huge amount of bias field in your image.
There's a trick in 3 steps to solve this problem.
1) Smooth your brain extracted image with fslmaths:
fslmaths exam_FAi_brain -s 20 smoothed_brain -odt
float
2) Smooth the mask of this brain:
fslmaths exam_FAi_brain -bin -s 20 smoothed_brain_mask
-odt float
3) Create your new "corrected" brain image:
fslmaths exam_FAi_brain -div smoothed_brain -mul
smoothed_brain_mask exam_FAi_brain_corrected -odt
float.
Once you've made sure that all the new corrected brain
images are fine, move them into a new directory
renaming them exam_FAi_brain, preprocess them again
with tbss_1_preproc and the appropriate scaling factor
then run fslvbm2 on them. Finally, move them back to
original FAi directory and the rest of the
non-problematic dataset.
Hope this makes sense,
Gwenaelle
--- "Thomas, Seth" <[log in to unmask]> a écrit
:
> Hello Gwenaelle,
>
> Thank you for inviting me to upload the file. The
> ID is 583445, and I have uploaded both GM_mod_merg
> and GM_mod_merg_s4 for comparison. It seems
> plausible that the smoothing kernel lowered the
> intensities, but I was surprised at the degree to
> which this happened, and am not sure whether the
> faintness of the output image means there was a
> problem at an earlier step or if this result is
> typical. For example, why is the minimum of the
> range (of the smoothed image) a negative number and
> not zero?
>
> If in fact these are ok, do you recommend that I
> change any attributes of the file itself (using
> fslmaths, for example) before going on to the final
> step of VBM analysis, or are we safe simply changing
> the brightness onscreen with FSLView?
>
> Thanks again!
>
> Seth
>
>
> -----Original Message-----
> From: FSL - FMRIB's Software Library on behalf of
> Gwenaëlle DOUAUD
> Sent: Tue 12/11/2007 6:29 AM
> To: [log in to unmask]
> Cc:
> Subject: [FSL] RE : [FSL] Concerns about VBM
> processing output
>
> Hi,
>
> range between 0 and 1000 seems reasonable. I'm not
> sure if your output from fslvbm_3 is fine or not as
> you have not adjusted the brightness. It should be
> ok
> though, but if you want to be sure, please upload
> your
> 4D GM_mod_merg on
> http://www.fmrib.ox.ac.uk/cgi-bin/upload.cgi and
> give
> us the ID...
>
> Hope this helps,
> Gwenaelle
>
>
> --- "Thomas, Seth" <[log in to unmask]> a
> écrit
> :
>
> >
> > We are performing VBM on a set of volumes and are
> > concerned about our results from the processing
> step
> > (fslvbm_3_proc). The attached file shows a
> snapshot
> > of GM_mod_merg_s2.5 as seen in FSLView. We are
> > unsure whether this output is reasonable, as our
> > template from the previous step appeared to have a
> > greater range of brightness in FSLView. Though we
> > were able to change the appearance of
> > GM_mod_merg_s2.5 in FSLView by simply entering the
> > arbitrary range 0 to 1000 in the min and max
> fields,
> > we realize that this does not change the image
> file
> > itself and we are not sure whether we should
> proceed
> > with these smoothed images as they are.
> >
> > Thank you in advance for any opinions anyone might
> > have about this as well as recommendations about
> > going on to the next step (in which the final
> > results are presented).
> >
> >
> >
>
----------------------------------------------------------
> > IMPORTANT WARNING: This email (and any
> attachments)
> > is only intended for the use of the person or
> entity
> > to which it is addressed, and may contain
> > information that is privileged and confidential.
> > You, the recipient, are obligated to maintain it
> in
> > a safe, secure and confidential manner.
> > Unauthorized redisclosure or failure to maintain
> > confidentiality may subject you to federal and
> state
> > penalties. If you are not the recipient, please
> > immediately notify us by return email, and delete
> > this message from your computer.
> >
>
----------------------------------------------------------
> >
> >
>
>
>
>
>
_____________________________________________________________________________
>
> Ne gardez plus qu'une seule adresse mail ! Copiez
> vos mails vers Yahoo! Mail http://mail.yahoo.fr
>
>
>
>
>
----------------------------------------------------------
> IMPORTANT WARNING: This email (and any attachments)
> is only intended for the use of the person or entity
> to which it is addressed, and may contain
> information that is privileged and confidential.
> You, the recipient, are obligated to maintain it in
> a safe, secure and confidential manner.
> Unauthorized redisclosure or failure to maintain
> confidentiality may subject you to federal and state
> penalties. If you are not the recipient, please
> immediately notify us by return email, and delete
> this message from your computer.
>
----------------------------------------------------------
>
_____________________________________________________________________________
Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail http://mail.yahoo.fr
|