Hi Donna,
The older version of caret_command works fine, in both preborder and
postborder, but when we need to adjust the landmarks using caret5's GUI,
we got this error:
Gtk-CRITICAL **: IA__gtk_widget_style_get: assertion `GTK_IS_WIDGET
(widget)' failed
The computer then pretty much freezes. It seems there's something
wrong with X-windows? We're running 64bit Linux (Ubuntu). Do you have
any suggestions? I couldn't find any obvious solution by searching the
web. The new caret5 GUI works just fine.
Thank you,
--ts
On 11/27/2013 05:18 PM, Donna Dierker wrote:
Good question. You could use different versions for the two steps, but you'd
have to report that in your methods, which could get messy. Running pre
border.sh with the older caret_command probably wouldn't take that long, but if
you tweaked your borders, it could overwrite them, depending on what you used
for the borderproj update string (e.g., updated).
To get around that, you could tar/zip just your tweaked borderproj files;
re-run preborder.sh; unzip the preserved tweaked borderproj; and run post
border.sh.
On Nov 27, 2013, at 2:01 PM, Taosheng Liu <[email protected]> wrote:
Thanks for the tip. Just another question, if we switch to the older version,
do you think we should re-run the preborder script as well? or we can pick up
from where preborder has ended and just run postborder with the older
caret_command?
Thank you,
--ts
On 11/27/2013 02:33 PM, Donna Dierker wrote:
I agree, the check_reg captures look great to me. The 144 cycles are
troubling. In your shoes, I'd revert.
On Nov 27, 2013, at 10:41 AM, Taosheng Liu <[email protected]> wrote:
Hi Donna,
Thanks for testing this, and for your fast response. Yes we're not getting a
core dump and it does finish, although with 144 cycles. Are you saying we
should switch back to the old caret for the time being? I mean, the check_reg
result seems fine, but maybe there's still something wrong with it?
Thank you!
--ts
On 11/27/2013 11:27 AM, Donna Dierker wrote:
Hi Taosheng,
I tried running postborder.sh on a dataset I have here, and I got a core dump
on the offending line:
+ caret_command -surface-sphere-multi-morph Human.SAIS_018.L.73730.spec
deformed_Human.SAIS_018.L.Midthickness_711-2B.mws.coord
Human.SAIS_018.L.SPHERICAL.73730.MWS.coord Human.SAIS_018.L.CLOSED.73730.topo
./PALS_B12.LR/postborder.sh: line 50: 27257 Segmentation fault (core
dumped) caret_command -surface-sphere-multi-morph $SPEC $FIDUCIAL_MWS
$SPHERE_MWS $TOPO
You're not getting a core dump, but something else is clearly going awry with
that line.
I tried backtracking to a June 2011 vintage caret, and it completed with no
problems. Here is the version I used:
http://brainvis.wustl.edu/pub/caret/caret_distribution_Linux64.v5.64.zip
login pub
password download
Here are the logs for the for the two versions of caret (bad and good):
Problem log:
http://brainmap.wustl.edu/pub/donna/US/MI/postborder.log
Good log:
http://brainvis.wustl.edu/pub/donna/US/MI/postborder.201106.log
I'm not sure when caret_command will be fixed, but I will pass the dataset and
details on to the developers.
Donna
On Nov 26, 2013, at 11:58 AM, Donna Dierker <[email protected]> wrote:
Hi Taosheng,
I am going to try postborder.sh on my new Linux box and see if it does the same
thing.
I don't see any candidate culprits in my ~/.caret5_preferences file.
No one else has reported this issue. More users are moving to the fs_LR
pipeline:
http://brainvis.wustl.edu/wiki/index.php/Caret:Operations/Freesurfer_to_fs_LR
There are trade-offs which are discussed here:
http://cercor.oxfordjournals.org/content/early/2011/11/02/cercor.bhr291.full.pdf+html
One big plus is that there is no border tweaking.
I'll let you know how my post border.sh trial comes out, or if I need sample
data. But I think I have plenty of sample data right here. I can't imagine
what might be unique about our data that would cause this.
Donna
On Nov 26, 2013, at 8:22 AM, Taosheng Liu <[email protected]> wrote:
Hi Donna,
Yes I can see that line of code in the script. Have you other others on this
list run this command lately? I wonder if you also see this many cycles. If
not, I wonder if there is some global setting I should set, or something is off
with my data. I can certainly supply the data if needed.
Thank you,
--Taosheng
On 11/25/2013 06:50 PM, Donna Dierker wrote:
Hmmm. Your results look great, but 144 cycles of spherical morphing is
certainly not normal. I think the line in the script that does this is this
one:
#MULTIRESOLUTION MORPHING
caret_command -surface-sphere-multi-morph $SPEC $FIDUCIAL_MWS $SPHERE_MWS
$TOPO
There are no parameters for this one; they are built in, apparently, so I'm not
sure what could be going on.
I'm stumped.
On Nov 25, 2013, at 4:27 PM, Taosheng Liu <[email protected]> wrote:
Hi Donna,
I've been using the preborder and postborder scripts to convert files from
FreeSurfer to Caret (hope you still remember I was involved in early testing of
these scripts). Things have been working great. I haven't done this for a
while, but recently I changed computer and installed the current version of
Caret and these scripts, and there seems to be some difference in how the code
works. I just want to make sure this is expected behavior.
Specifically, when we run postborder script, it takes much longer. It seems
either the script or caret_command is doing more iterations, because we now see
files like this after the script finishes:
Human.FS040.L.SPHERICAL.73730.MWS.SPHERE_CYCLE144.coord
There're a lot of files like this, with consecutive names from CYCLE1 to CYCLE
144, whereas previously it only goes from CYCLE1 to CYCLE4. My question is: is
this expected behavior, or is there something wrong in how we ran the scripts?
I've posted the result from the check_reg here
http://psychology.msu.edu/liulab/files/html_out/views.html
They look just like before, but I'd like to confirm whether this is normal.
Thank you very much,
--Taosheng
--
Taosheng Liu
Assistant Professor
Department of Psychology
Michigan State University
316 Physics Road
East Lansing, MI 48824
PH: 517.432.6694
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
--
Taosheng Liu
Assistant Professor
Department of Psychology
Michigan State University
316 Physics Road
East Lansing, MI 48824
PH: 517.432.6694
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
--
Taosheng Liu
Assistant Professor
Department of Psychology
Michigan State University
316 Physics Road
East Lansing, MI 48824
PH: 517.432.6694
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
--
Taosheng Liu
Assistant Professor
Department of Psychology
Michigan State University
316 Physics Road
East Lansing, MI 48824
PH: 517.432.6694
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
--
Taosheng Liu
Assistant Professor
Department of Psychology
Michigan State University
316 Physics Road
East Lansing, MI 48824
PH: 517.432.6694
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
