I got the following output after running the second stage of `rebuild_gca_atlas.csh', which starts by training on all segmented subjects. The first stage (which performs a Tal reg on the first subject listed in env var $SUBJECTS) finished without problems:
echo "$train using all subjects, using ${M3D_ONE}, producing ${GCA}..." set cmd=($train -prior_spacing 2 -node_spacing 4 -mask ${MASK_VOL}) set cmd=($cmd -parc_dir ${SEG_VOL} -xform ${M3D_ONE} -T1 ${T1_VOL} -check) set cmd=($cmd ${SUBJECTS} ${GCA})
--------- Execution stopped after this step: --------- (STDOUT) computing covariances for subject A59, 1 of 4... reading volume /Users/jsm/work/subjects/A59/mri/brain.mgz for masking... gunzip -c /Users/jsm/work/subjects/A59/mri/transforms/talairach_one.m3z GCAupdateNodeCovariance(10, 35, 23, 4): could not find label ---------
Here `$SUBJECTS' was set to 'A59 A62 f18 f20', with "A59" being the first subject.
Thanks in advance, Jason
Hello,
tkmedit and tksurfer fail to start because of bus errors. The problem seems to happen at random times, but never when launching one of these programs for the first time after power-cycling the machine.
(from `uname -a') Darwin localhost 9.5.0 Darwin Kernal Version 9.5.0: Wed Sep 3 11:29 PDT 2008; root:xnu-1228.7.58~1/RELEASE_I386 i386
% tkmedit $SUBJECT brainmask.mgz Bus Error
Is this a known issue? Thanks in advance.
Jason
Jason,
I'm not aware of this issue. What distribution do you have? Type 'bugr' to get this info.
Nick
On Fri, 2008-11-07 at 13:42 -0800, Jason S. Minamora wrote:
Hello,
tkmedit and tksurfer fail to start because of bus errors. The problem seems to happen at random times, but never when launching one of these programs for the first time after power-cycling the machine.
(from `uname -a') Darwin localhost 9.5.0 Darwin Kernal Version 9.5.0: Wed Sep 3 11:29 PDT 2008; root:xnu-1228.7.58~1/RELEASE_I386 i386
% tkmedit $SUBJECT brainmask.mgz Bus Error
Is this a known issue? Thanks in advance.
Jason
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
mri_ca_train (called from rebuild_gca_atlas.csh) terminated after printing the following to screen:
----------- Log file is /Users/jsm/work/subjects/rebuild_gca_atlas_2008-11-11.log mri_ca_normalize each subject using its seg_edited.mgz, producing norm.mgz... using MR volume /Users/jsm/womri_ca_train using one subject: f18, producing /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca... pbsubmit: Command not found. erate control points... reading 1 input volumes reading atlas from 'noatlas'... reading input volume from /Users/jsm/work/subjects/f18/mri/nu.mgz... bias field = 1.403 +- 0.096 361 of 19551 control points discarded bias field = 0.971 +- 0.051 834 of 34697 control points discarded writing normalized volume to /Users/jsm/work/subjects/f18/mri/norm.mgz... freeing GCA...done. normalization took 2 minutes and 29 seconds. mri_ca_train using one subject: f18, producing /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca... reading input 0: /Users/jsm/work/subjects/f18/mri/norm.mgz spacing priors every 2.0 mm spacing nodes every 8.0 mm using MR volume brain.mgz to mask input volume... reading segmentation from subject's mri/seg_edited.mgz directory reading xform from talairach_man.xfm reading T1 data from subject's mri/norm.mgz directory will conduct sanity-check of labels... training on 1 subject and writing results to /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca *************************************************************************** processing subject f18, 1 of 1... reading volume /Users/jsm/work/subjects/f18/mri/brain.mgz for masking... INFO: original RAS-to-RAS transform 1.076 0.007 0.006 -3.332; -0.094 1.129 0.384 -27.287; 0.003 -0.370 1.069 -0.379; 0.000 0.000 0.000 1.000; INFO: modified VOX-to-VOX transform 1.076 0.006 -0.007 -10.648; 0.003 1.069 0.370 -70.266; 0.094 -0.384 1.129 33.189; 0.000 0.000 0.000 1.000; checking labels in subject f18... ERROR: Right_Hippocampus: 104 138 97, tal x=26.245338, y=-27.008644, *** z=14.712428 > 13 *** ERROR: Right_Hippocampus: 104 139 96, tal x=26.232159, y=-28.505787, *** z=14.143823 > 13 *** ERROR: Right_Hippocampus: 104 139 97, tal x=26.239115, y=-27.429529, *** z=13.748710 > 13 *** ERROR: Right_Hippocampus: 105 138 97, tal x=25.179642, y=-26.918175, *** z=14.705179 > 13 *** ERROR: Right_Hippocampus: 105 139 97, tal x=25.173420, y=-27.339060, *** z=13.741462 > 13 *** ERROR: Right_Hippocampus: 106 139 97, tal x=24.107723, y=-27.248587, *** z=13.734215 > 13 *** ERROR: Right_Hippocampus: 110 139 96, tal x=19.837978, y=-27.962954, *** z=14.100336 > 13 *** (...) ERROR: Left_Hippocampus: 155 141 92, tal x=-28.158634, y=-29.038506, *** z=13.427198 > 13 *** ERROR: Left_Hippocampus: 155 141 93, tal x=-28.151674, y=-27.962250, *** z=13.032085 > 13 *** max_xtal_l_hippo = -8.9 max_xtal_l_caudate = -3.3 max_xtal_l_amygdala = -13.1 max_xtal_l_putamen = -11.9 max_xtal_l_pallidum = -10.9 min_xtal_r_hippo = 8.1 min_xtal_r_caudate = 5.2 min_xtal_r_amygdala = 11.4 min_xtal_r_putamen = 11.5 min_xtal_r_pallidum = 12.5 ERROR: mri_ca_train: possible bad training data! subject: /Users/jsm/work/subjects/f18
ERROR: mri_ca_train check found 1 subjects with bad labels!
978 prior holes filled filling holes in the GCA... 38886 classifiers: 1.2 per node, 2.82 in brain (0 holes filled)
-------------
Is there a way to override the default thresholds?
Jason
Jason,
There is no way to adjust the thresholds. You can eliminate the checking by editing the rebuild_gca_atlas file and removing the '-check' flag from the mri_ca_train command (in two places in the script).
The checking performed by mri_ca_train is intended to catch gross errors (like a left_hippocampus label in the right hemi). In your case, i would just look at each error to look for gross errors. I don't see any in your output though.
Nick
On Tue, 2008-11-11 at 14:36 -0800, Jason S. Minamora wrote:
mri_ca_train (called from rebuild_gca_atlas.csh) terminated after printing the following to screen:
Log file is /Users/jsm/work/subjects/rebuild_gca_atlas_2008-11-11.log mri_ca_normalize each subject using its seg_edited.mgz, producing norm.mgz... using MR volume /Users/jsm/womri_ca_train using one subject: f18, producing /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca... pbsubmit: Command not found. erate control points... reading 1 input volumes reading atlas from 'noatlas'... reading input volume from /Users/jsm/work/subjects/f18/mri/nu.mgz... bias field = 1.403 +- 0.096 361 of 19551 control points discarded bias field = 0.971 +- 0.051 834 of 34697 control points discarded writing normalized volume to /Users/jsm/work/subjects/f18/mri/norm.mgz... freeing GCA...done. normalization took 2 minutes and 29 seconds. mri_ca_train using one subject: f18, producing /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca... reading input 0: /Users/jsm/work/subjects/f18/mri/norm.mgz spacing priors every 2.0 mm spacing nodes every 8.0 mm using MR volume brain.mgz to mask input volume... reading segmentation from subject's mri/seg_edited.mgz directory reading xform from talairach_man.xfm reading T1 data from subject's mri/norm.mgz directory will conduct sanity-check of labels... training on 1 subject and writing results to /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca
processing subject f18, 1 of 1... reading volume /Users/jsm/work/subjects/f18/mri/brain.mgz for masking... INFO: original RAS-to-RAS transform 1.076 0.007 0.006 -3.332; -0.094 1.129 0.384 -27.287; 0.003 -0.370 1.069 -0.379; 0.000 0.000 0.000 1.000; INFO: modified VOX-to-VOX transform 1.076 0.006 -0.007 -10.648; 0.003 1.069 0.370 -70.266; 0.094 -0.384 1.129 33.189; 0.000 0.000 0.000 1.000; checking labels in subject f18... ERROR: Right_Hippocampus: 104 138 97, tal x=26.245338, y=-27.008644, *** z=14.712428 > 13 *** ERROR: Right_Hippocampus: 104 139 96, tal x=26.232159, y=-28.505787, *** z=14.143823 > 13 *** ERROR: Right_Hippocampus: 104 139 97, tal x=26.239115, y=-27.429529, *** z=13.748710 > 13 *** ERROR: Right_Hippocampus: 105 138 97, tal x=25.179642, y=-26.918175, *** z=14.705179 > 13 *** ERROR: Right_Hippocampus: 105 139 97, tal x=25.173420, y=-27.339060, *** z=13.741462 > 13 *** ERROR: Right_Hippocampus: 106 139 97, tal x=24.107723, y=-27.248587, *** z=13.734215 > 13 *** ERROR: Right_Hippocampus: 110 139 96, tal x=19.837978, y=-27.962954, *** z=14.100336 > 13 *** (...) ERROR: Left_Hippocampus: 155 141 92, tal x=-28.158634, y=-29.038506, *** z=13.427198 > 13 *** ERROR: Left_Hippocampus: 155 141 93, tal x=-28.151674, y=-27.962250, *** z=13.032085 > 13 *** max_xtal_l_hippo = -8.9 max_xtal_l_caudate = -3.3 max_xtal_l_amygdala = -13.1 max_xtal_l_putamen = -11.9 max_xtal_l_pallidum = -10.9 min_xtal_r_hippo = 8.1 min_xtal_r_caudate = 5.2 min_xtal_r_amygdala = 11.4 min_xtal_r_putamen = 11.5 min_xtal_r_pallidum = 12.5 ERROR: mri_ca_train: possible bad training data! subject: /Users/jsm/work/subjects/f18
ERROR: mri_ca_train check found 1 subjects with bad labels!
978 prior holes filled filling holes in the GCA... 38886 classifiers: 1.2 per node, 2.82 in brain (0 holes filled)
Is there a way to override the default thresholds?
Jason _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Hi Jason,
those are sanity checks we put in to make sure that we didn't accidentally have say right hippocampal labels in the left hemi. Check the manually labeled subjects and see if they have the wrong hemi label.
cheers, Bruce
On Tue, 11 Nov 2008, Jason S. Minamora wrote:
mri_ca_train (called from rebuild_gca_atlas.csh) terminated after printing the following to screen:
Log file is /Users/jsm/work/subjects/rebuild_gca_atlas_2008-11-11.log mri_ca_normalize each subject using its seg_edited.mgz, producing norm.mgz... using MR volume /Users/jsm/womri_ca_train using one subject: f18, producing /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca... pbsubmit: Command not found. erate control points... reading 1 input volumes reading atlas from 'noatlas'... reading input volume from /Users/jsm/work/subjects/f18/mri/nu.mgz... bias field = 1.403 +- 0.096 361 of 19551 control points discarded bias field = 0.971 +- 0.051 834 of 34697 control points discarded writing normalized volume to /Users/jsm/work/subjects/f18/mri/norm.mgz... freeing GCA...done. normalization took 2 minutes and 29 seconds. mri_ca_train using one subject: f18, producing /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca... reading input 0: /Users/jsm/work/subjects/f18/mri/norm.mgz spacing priors every 2.0 mm spacing nodes every 8.0 mm using MR volume brain.mgz to mask input volume... reading segmentation from subject's mri/seg_edited.mgz directory reading xform from talairach_man.xfm reading T1 data from subject's mri/norm.mgz directory will conduct sanity-check of labels... training on 1 subject and writing results to /Users/jsm/work/subjects/average/RB_one_2008-11-11.gca
processing subject f18, 1 of 1... reading volume /Users/jsm/work/subjects/f18/mri/brain.mgz for masking... INFO: original RAS-to-RAS transform 1.076 0.007 0.006 -3.332; -0.094 1.129 0.384 -27.287; 0.003 -0.370 1.069 -0.379; 0.000 0.000 0.000 1.000; INFO: modified VOX-to-VOX transform 1.076 0.006 -0.007 -10.648; 0.003 1.069 0.370 -70.266; 0.094 -0.384 1.129 33.189; 0.000 0.000 0.000 1.000; checking labels in subject f18... ERROR: Right_Hippocampus: 104 138 97, tal x=26.245338, y=-27.008644, *** z=14.712428 > 13 *** ERROR: Right_Hippocampus: 104 139 96, tal x=26.232159, y=-28.505787, *** z=14.143823 > 13 *** ERROR: Right_Hippocampus: 104 139 97, tal x=26.239115, y=-27.429529, *** z=13.748710 > 13 *** ERROR: Right_Hippocampus: 105 138 97, tal x=25.179642, y=-26.918175, *** z=14.705179 > 13 *** ERROR: Right_Hippocampus: 105 139 97, tal x=25.173420, y=-27.339060, *** z=13.741462 > 13 *** ERROR: Right_Hippocampus: 106 139 97, tal x=24.107723, y=-27.248587, *** z=13.734215 > 13 *** ERROR: Right_Hippocampus: 110 139 96, tal x=19.837978, y=-27.962954, *** z=14.100336 > 13 *** (...) ERROR: Left_Hippocampus: 155 141 92, tal x=-28.158634, y=-29.038506, *** z=13.427198 > 13 *** ERROR: Left_Hippocampus: 155 141 93, tal x=-28.151674, y=-27.962250, *** z=13.032085 > 13 *** max_xtal_l_hippo = -8.9 max_xtal_l_caudate = -3.3 max_xtal_l_amygdala = -13.1 max_xtal_l_putamen = -11.9 max_xtal_l_pallidum = -10.9 min_xtal_r_hippo = 8.1 min_xtal_r_caudate = 5.2 min_xtal_r_amygdala = 11.4 min_xtal_r_putamen = 11.5 min_xtal_r_pallidum = 12.5 ERROR: mri_ca_train: possible bad training data! subject: /Users/jsm/work/subjects/f18
ERROR: mri_ca_train check found 1 subjects with bad labels!
978 prior holes filled filling holes in the GCA... 38886 classifiers: 1.2 per node, 2.82 in brain (0 holes filled)
Is there a way to override the default thresholds?
Jason _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Hi,
I'm having trouble getting the automatic segmentation algorithm to bind to the wm and pial surfaces: several areas are excluded, and I'm pretty sure it's because of poor contrast. Is there a way to save the brightness/contrast adjustments made in tkmedit? I've tried View > Configure > Brightness / Contrast to discover the excluded areas, and would like to actually save these settings to the volume itself. Is that possible in tkmedit or anywhere else?
Cheers, Jason
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
Jason S. Minamora -
Nick Schmansky