Hi all,
I posted this problem earlier, but did not get an answer, so I am posting it
again, sorry!
---
recon-all runs smoothly for all my subjects but one. It throws the error
"mri_watershed Error: indices out of bounds" and finishes (see
https://www.dropbox.com/s/4del0uy8g1qw5az/recon-all.
Hi all,
recon-all runs smoothly for all my subjects but one. It throws the error
"mri_watershed Error: indices out of bounds" and finishes (see
https://www.dropbox.com/s/4del0uy8g1qw5az/recon-all.log). The image quality
does not seem to be different...
The error has been discussed on the list
Hi Tim
that's a new error - congratulations! If you upload the dataset I'll take a
look, but I suspect if you add some control points it will work fine.
cheers
Bruce
On Wed, 11 Jan 2012, Holden, Timothy
wrote:
Hi FS experts,
I?m encountering an error in the Skull Strip routine with
sorry, it doesn't sound like this is the problem. You could try putting
some seed points in the white matter for the watershed. Not sure what the
interface is for that. You should also check to make sure that the atlas
transform it is using is reasonable.
On Wed, 7 Oct
2009, Gonzalo Rojas wro
check to see if the intensity normalization worked (does the T1.mgz white
matter have intensities of 110 most places)
On Wed, 7 Oct 2009, Gonzalo
Rojas wrote:
> Hi:
>
> I got the following error... How can I solve it ?...
>
> recon-all.log
>
Hi:
Ok... I test it with the command that you told me, and I added the
option "wsthresh 35"... Then I got a good skull stripped image...
Sincerely,
Gonzalo Rojas Costa
Department of Radiology
Las Condes Clinic
Lo Fontecilla 441, Las Condes, Santiago, Chile.
Tel: 56-2-2105170
Cel: 56-9-97
another thing to try is include the flag -no-wsgcaatlas flag, like this:
recon-all -s subjid -skull-strip -no-wsgcaatlas
this bypasses usage of the atlas.
if that fails, you can send me the 001.mgz input file to the filedrop:
https://www.nmr.mgh.harvard.edu/facility/filedrop/index.html
n.
On
Hi:
How can I put the control points in the nu.mgz image ?...
Sincerely
Gonzalo Rojas Costa
Department of Radiology
Las Condes Clinic
Lo Fontecilla 441, Las Condes, Santiago, Chile.
Tel: 56-2-2105170
Cel: 56-9-97771785
www.clc.cl
Nick Schmansky escribió:
> check if the whitematter in th
Hi:
Yes... the occipital white matter has value 110 in general... The
frontal white matter has voxels with values 94 to 117 aprox...
Sincerely,
Gonzalo Rojas Costa
Department of Radiology
Las Condes Clinic
Lo Fontecilla 441, Las Condes, Santiago, Chile.
Tel: 56-2-2105170
Cel: 56-9-977717
check if the whitematter in the file T1.mgz is around value 110. if
not, then put some control points in the white matter in nu.mgz and
re-run recon-all.
n.
On Wed, 2009-10-07 at 15:34 -0400, Gonzalo Rojas wrote:
> Hi:
>
>I got the following error... How can I solve it ?...
>
> recon-all.l
Hi:
I got the following error... How can I solve it ?...
recon-all.log
-
mri_watershed -T1 -brain_atlas
/usr/local/freesurfer/average/RB_all_withskull_2008-03-26.gca
transforms/talairach_with_skull.lta T1.mgz brainmask.auto.mg
Vina,
Can you check the FOV value for the orig.mgz file? That is:
cd subj/mri
mri_info orig.mgz
Is it greater than 256? Are the dimensions >256?
I have just recently been looking into this error for someone else, and
the source of the problem was improper handling by mri_convert of data
tha
Did you explicitly check the quality of talairach_with_skull.lta?
Also, what version of FS are you using?
We previously had encountered that error in 5 brains processed under FS
4.0.2, and all five of those had errant talairach_with_skull.lta
transforms. Generation of talairach_with_skull.lta wa
hmmm, that's interesting. If you send us the data we'll take a look. In
any case, you can run, if you create an orig_stripped.mgz you can run:
mri_mask T1.mgz orig_stripped.mgz brain.mgz
then you should be able to run from after the skull stripping forward
(check the recon-all help. You'll ne
Hi all,
I get this error for one subject: mri_watershed Error:
GLOBAL region of the brain empty !
I check the talairach transformation and it looked good. I also tried
just scripting the orig.mgz and I get a wonderful strip. How can I move
forward with the recon-all process?
Thanks,
Vina
__
15 matches
Mail list logo
