Hi, Freesurfers I am getting some interesting hippocampal surfaces with the following steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement volumes) 3- a custom OCTAVE script to segment the concatenated volume from step 2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5
Using freeview I got a screenshot showing both hippocampi segmented (hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6 was completed as well in 13 out 16 subjects.
The only problem was during left hippocampal body smoothing in 3 subjects (all the other surfaces from all 16 subjects in were smoothed with no problem). For these 3 surfaces, mris_smooth simply freezes, with no error message. I must press control + C to quit the program. During the freezing I noticed high CPU use. It follows the exactly command I used and the screen output.
mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody using Gaussian curvature smoothing with norm 5.00 with 2 smooth steps averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------------------------------ ------------------------------------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------------------------------ ------------------------------------------------------
If I supress the gaussian parameters and run just something like: mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody
I got the smoothed surface with no errors but the final quality is not as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you can check it (surfLeftHippobody). Intersting if I use a surface with all left hippocampus (including head, problematic body and tail and) smoothing completes without error Thank you very much for any help. Marcos
how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote:
Hi, Freesurfers I am getting some interesting hippocampal surfaces with the following steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement volumes) 3- a custom OCTAVE script to segment the concatenated volume from step 2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5
Using freeview I got a screenshot showing both hippocampi segmented (hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6 was completed as well in 13 out 16 subjects.
The only problem was during left hippocampal body smoothing in 3 subjects (all the other surfaces from all 16 subjects in were smoothed with no problem). For these 3 surfaces, mris_smooth simply freezes, with no error message. I must press control + C to quit the program. During the freezing I noticed high CPU use. It follows the exactly command I used and the screen output.
mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody using Gaussian curvature smoothing with norm 5.00 with 2 smooth steps averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations...
------------------ pass 1 (num=2853) ------------------
If I supress the gaussian parameters and run just something like: mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody
I got the smoothed surface with no errors but the final quality is not as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you can check it (surfLeftHippobody). Intersting if I use a surface with all left hippocampus (including head, problematic body and tail and) smoothing completes without error Thank you very much for any help. Marcos
Hi, Bruce. Thank you for your fast answer. I used a script and It took less than 15 min to process all files from all subjects, except the 3 files (surfaces) mris_smooth seems to enter a loop. When I tried to smooth surfLeftHippobody I waited for more than 4 hours. Comparing to the total time I needed to process all files I concluded something was wrong and esed control + C to interrupt the process. I tested on 2 machines (mine is running CentOS 6 and the second is a Mac, both using Freesurfer 5.3) and got the same result. Including they seem to stop at the same point. Do you think it could need more time to finnish processing? Alternatively, as I can use mris_smooth to process the whole left hippocampus (the problem only appears when processing the body segment), I guess I could get a similar and perhaps better result if I could make a label from each of the binary masks of the 3 segments (head, body, tail), merge them in an annot file and finally load the whole surface and the annot file so I get the 3 segments discriminated by color. I found on Freesurfer Wiki how to merge the labels into an annot file but not how to make the labels from the binary masks. Thank you for your attention.
2013/10/4 Bruce Fischl fischl@nmr.mgh.harvard.edu
how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote:
Hi, Freesurfers I am getting some interesting hippocampal surfaces with the following steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement volumes) 3- a custom OCTAVE script to segment the concatenated volume from step 2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5
Using freeview I got a screenshot showing both hippocampi segmented (hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6 was completed as well in 13 out 16 subjects.
The only problem was during left hippocampal body smoothing in 3 subjects (all the other surfaces from all 16 subjects in were smoothed with no problem). For these 3 surfaces, mris_smooth simply freezes, with no error message. I must press control + C to quit the program. During the freezing I noticed high CPU use. It follows the exactly command I used and the screen output.
mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody using Gaussian curvature smoothing with norm 5.00 with 2 smooth steps averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------**------------------------ ------------------------------**------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------**------------------------ ------------------------------**------------------------
If I supress the gaussian parameters and run just something like: mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody
I got the smoothed surface with no errors but the final quality is not as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you can check it (surfLeftHippobody). Intersting if I use a surface with all left hippocampus (including head, problematic body and tail and) smoothing completes without error Thank you very much for any help. Marcos
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/**compliancelinehttp://www.partners.org/complianceline. If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
hmmm, I guess you should upload the subject so that we can track it down On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce. Thank you for your fast answer. I used a script and It took less than 15 min to process all files from all subjects, except the 3 files (surfaces) mris_smooth seems to enter a loop. When I tried to smooth surfLeftHippobody I waited for more than 4 hours. Comparing to the total time I needed to process all files I concluded something was wrong and esed control + C to interrupt the process. I tested on 2 machines (mine is running CentOS 6 and the second is a Mac, both using Freesurfer 5.3) and got the same result. Including they seem to stop at the same point. Do you think it could need more time to finnish processing? Alternatively, as I can use mris_smooth to process the whole left hippocampus (the problem only appears when processing the body segment), I guess I could get a similar and perhaps better result if I could make a label from each of the binary masks of the 3 segments (head, body, tail), merge them in an annot file and finally load the whole surface and the annot file so I get the 3 segments discriminated by color. I found on Freesurfer Wiki how to merge the labels into an annot file but not how to make the labels from the binary masks. Thank you for your attention.
2013/10/4 Bruce Fischl fischl@nmr.mgh.harvard.edu how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote: Hi, Freesurfers I am getting some interesting hippocampal surfaces with the following steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement volumes) 3- a custom OCTAVE script to segment the concatenated volume from step 2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5 Using freeview I got a screenshot showing both hippocampi segmented (hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6 was completed as well in 13 out 16 subjects. The only problem was during left hippocampal body smoothing in 3 subjects (all the other surfaces from all 16 subjects in were smoothed with no problem). For these 3 surfaces, mris_smooth simply freezes, with no error message. I must press control + C to quit the program. During the freezing I noticed high CPU use. It follows the exactly command I used and the screen output. mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody using Gaussian curvature smoothing with norm 5.00 with 2 smooth steps averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------------------------------ ------------------------------------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------------------------------ ------------------------------------------------------ If I supress the gaussian parameters and run just something like: mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody I got the smoothed surface with no errors but the final quality is not as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you can check it (surfLeftHippobody). Intersting if I use a surface with all left hippocampus (including head, problematic body and tail and) smoothing completes without error Thank you very much for any help. MarcosThe information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Hi, Bruce I proceeded as explained at http://surfer.nmr.mgh.harvard.edu/fswiki/FtpFileExchange . Using ftp method, it started at good rate but after I transferred about 3/4 of the file the transfer rate decreased to some bytes (not Kbytes) per second. Tried using curl -T and it seemed to complete after about 50 min but I got this message on terminal: DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN"> <html><head> <title>405 Method Not Allowed</title> </head><body> <h1>Method Not Allowed</h1> <p>The requested method PUT is not allowed for the URL /transfer/incoming.</p> <hr> <address>Apache/2.2.15 (CentOS) Server at surfer.nmr.mgh.harvard.edu Port 80</address> </body></html>
So I am not sure you received the file. You should find a suj46.zip corresponding to the complete folder from my subject 46. You will find the usual structure created by recon-all. At mri folder you will find all the usual volumes, the posterior* files generated by hippo-subfield processing and the files I created on my analysis. Please, could you check if you really received the file? If not, I will try again using plain ftp. Thank you again for your help. Marcos. Em Seg, 2013-10-07 às 09:21 -0400, Bruce Fischl escreveu:
hmmm, I guess you should upload the subject so that we can track it down On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce. Thank you for your fast answer. I used a script and It took less than 15 min to process all files from all subjects, except the 3 files (surfaces) mris_smooth seems to enter a loop. When I tried to smooth surfLeftHippobody I waited for more than 4 hours. Comparing to the total time I needed to process all files I concluded something was wrong and esed control + C to interrupt the process. I tested on 2 machines (mine is running CentOS 6 and the second is a Mac, both using Freesurfer 5.3) and got the same result. Including they seem to stop at the same point. Do you think it could need more time to finnish processing? Alternatively, as I can use mris_smooth to process the whole left hippocampus (the problem only appears when processing the body segment), I guess I could get a similar and perhaps better result if I could make a label from each of the binary masks of the 3 segments (head, body, tail), merge them in an annot file and finally load the whole surface and the annot file so I get the 3 segments discriminated by color. I found on Freesurfer Wiki how to merge the labels into an annot file but not how to make the labels from the binary masks. Thank you for your attention.
2013/10/4 Bruce Fischl fischl@nmr.mgh.harvard.edu how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote: Hi, Freesurfers I am getting some interesting hippocampal surfaces with the following steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement volumes) 3- a custom OCTAVE script to segment the concatenated volume from step 2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5 Using freeview I got a screenshot showing both hippocampi segmented (hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6 was completed as well in 13 out 16 subjects. The only problem was during left hippocampal body smoothing in 3 subjects (all the other surfaces from all 16 subjects in were smoothed with no problem). For these 3 surfaces, mris_smooth simply freezes, with no error message. I must press control + C to quit the program. During the freezing I noticed high CPU use. It follows the exactly command I used and the screen output. mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody using Gaussian curvature smoothing with norm 5.00 with 2 smooth steps averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------------------------------ ------------------------------------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------------------------------ ------------------------------------------------------ If I supress the gaussian parameters and run just something like: mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody I got the smoothed surface with no errors but the final quality is not as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you can check it (surfLeftHippobody). Intersting if I use a surface with all left hippocampus (including head, problematic body and tail and) smoothing completes without error Thank you very much for any help. MarcosThe information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
sorry, I don't think it completed as unzip can't find the end of directory mark in the file. Try our filedrop instead:
https://gate.nmr.mgh.harvard.edu/filedrop2/
cheers Bruce
On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce I proceeded as explained at http://surfer.nmr.mgh.harvard.edu/fswiki/FtpFileExchange . Using ftp method, it started at good rate but after I transferred about 3/4 of the file the transfer rate decreased to some bytes (not Kbytes) per second. Tried using curl -T and it seemed to complete after about 50 min but I got this message on terminal: DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head> <title>405 Method Not Allowed</title> </head><body> <h1>Method Not Allowed</h1> <p>The requested method PUT is not allowed for the URL /transfer/incoming.</p> <hr> <address>Apache/2.2.15 (CentOS) Server at surfer.nmr.mgh.harvard.edu Port 80</address> </body></html>
So I am not sure you received the file. You should find a suj46.zip corresponding to the complete folder from my subject 46. You will find the usual structure created by recon-all. At mri folder you will find all the usual volumes, the posterior* files generated by hippo-subfield processing and the files I created on my analysis. Please, could you check if you really received the file? If not, I will try again using plain ftp. Thank you again for your help. Marcos. Em Seg, 2013-10-07 às 09:21 -0400, Bruce Fischl escreveu:
hmmm, I guess you should upload the subject so that we can track it down On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce. Thank you for your fast answer. I used a script and It took less than 15 min to process all files from all
subjects, except
the 3 files (surfaces) mris_smooth seems to enter a loop. When I tried to
smooth
surfLeftHippobody I waited for more than 4 hours. Comparing to the total t
ime I needed to
process all files I concluded something was wrong and esed control + C to
interrupt the
process. I tested on 2 machines (mine is running CentOS 6 and the second i
s a Mac, both using
Freesurfer 5.3) and got the same result. Including they seem to stop at th
e same point. Do you
think it could need more time to finnish processing? Alternatively, as I can use mris_smooth to process the whole left hippocam
pus (the problem
only appears when processing the body segment), I guess I could get a simi
lar and perhaps
better result if I could make a label from each of the binary masks of the
3 segments (head,
body, tail), merge them in an annot file and finally load the whole surfac
e and the annot file
so I get the 3 segments discriminated by color. I found on Freesurfer Wiki
how to merge the
labels into an annot file but not how to make the labels from the binary m
asks.
Thank you for your attention.
2013/10/4 Bruce Fischl fischl@nmr.mgh.harvard.edu how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote:
Hi, Freesurfers I am getting some interesting hippocampal surfaces with the fo
llowing
steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement v
olumes)
3- a custom OCTAVE script to segment the concatenated volume f
rom step
2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5
Using freeview I got a screenshot showing both hippocampi segm
ented
(hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6
was
completed as well in 13 out 16 subjects.
The only problem was during left hippocampal body smoothing in
3
subjects (all the other surfaces from all 16 subjects in were smoothed
with no
problem). For these 3 surfaces, mris_smooth simply freezes, wi
th no
error message. I must press control + C to quit the program. During
the
freezing I noticed high CPU use. It follows the exactly command I used and the screen output.
mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHipp
oBody
using Gaussian curvature smoothing with norm 5.00 with 2 smoot
h steps
averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------------------------------ ------------------------------------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------------------------------ ------------------------------------------------------
If I supress the gaussian parameters and run just something li
ke:
mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody
I got the smoothed surface with no errors but the final qualit
y is not
as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you can
check
it (surfLeftHippobody). Intersting if I use a surface with all le
ft
hippocampus (including head, problematic body and tail and) smoothing comp
letes
without error Thank you very much for any help. Marcos
The information in this e-mail is intended only for the person to whom it
is
addressed. If you believe this e-mail was sent to you in error and the e-m
ail
contains patient information, please contact the Partners Compliance HelpL
ine at
http://www.partners.org/complianceline . If the e-mail was sent to you in
error
but does not contain patient information, please contact the sender and pr
operly
dispose of the e-mail.
Hi, Bruce Filedrop worked great and fast! Thank you for the tip. It follows the bath I run for each subject: mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippo lh.surfHippo mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoHead lh.surfHippoHead mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoTail lh.surfHippoTail mris_smooth -nw -g 5 2 -a 1 -n 1 surfRightHippo rh.surfHippo mris_smooth -nw -g 5 2 -a 1 -n 1 surfRightHippoHead rh.surfHippoHead mris_smooth -nw -g 5 2 -a 1 -n 1 surfRightHippoBody rh.surfHippoBody mris_smooth -nw -g 5 2 -a 1 -n 1 surfRightHippoTail rh.surfHippoTail
The only file with the problem I reported is surfLeftHippoBody (3rd line). If I could generate an annotation file from surfLeftHippoHead, surfLeftHippoBody and surfLeftHippoTail or the binary masks I used mri_tessellate to make these surfaces it would be a good alternative. Perhaps, even better because I would get a faster processing (smoothing 2 surfaces, surfLeftHippo and surfRightHippo, per subject instead of 8). Thank you again. Em Seg, 2013-10-07 às 20:47 -0400, Bruce Fischl escreveu:
sorry, I don't think it completed as unzip can't find the end of directory mark in the file. Try our filedrop instead:
https://gate.nmr.mgh.harvard.edu/filedrop2/
cheers Bruce
On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce I proceeded as explained at http://surfer.nmr.mgh.harvard.edu/fswiki/FtpFileExchange . Using ftp method, it started at good rate but after I transferred about 3/4 of the file the transfer rate decreased to some bytes (not Kbytes) per second. Tried using curl -T and it seemed to complete after about 50 min but I got this message on terminal: DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head> <title>405 Method Not Allowed</title> </head><body> <h1>Method Not Allowed</h1> <p>The requested method PUT is not allowed for the URL /transfer/incoming.</p> <hr> <address>Apache/2.2.15 (CentOS) Server at surfer.nmr.mgh.harvard.edu Port 80</address> </body></html>
So I am not sure you received the file. You should find a suj46.zip corresponding to the complete folder from my subject 46. You will find the usual structure created by recon-all. At mri folder you will find all the usual volumes, the posterior* files generated by hippo-subfield processing and the files I created on my analysis. Please, could you check if you really received the file? If not, I will try again using plain ftp. Thank you again for your help. Marcos. Em Seg, 2013-10-07 às 09:21 -0400, Bruce Fischl escreveu:
hmmm, I guess you should upload the subject so that we can track it down On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce. Thank you for your fast answer. I used a script and It took less than 15 min to process all files from all
subjects, except
the 3 files (surfaces) mris_smooth seems to enter a loop. When I tried to
smooth
surfLeftHippobody I waited for more than 4 hours. Comparing to the total t
ime I needed to
process all files I concluded something was wrong and esed control + C to
interrupt the
process. I tested on 2 machines (mine is running CentOS 6 and the second i
s a Mac, both using
Freesurfer 5.3) and got the same result. Including they seem to stop at th
e same point. Do you
think it could need more time to finnish processing? Alternatively, as I can use mris_smooth to process the whole left hippocam
pus (the problem
only appears when processing the body segment), I guess I could get a simi
lar and perhaps
better result if I could make a label from each of the binary masks of the
3 segments (head,
body, tail), merge them in an annot file and finally load the whole surfac
e and the annot file
so I get the 3 segments discriminated by color. I found on Freesurfer Wiki
how to merge the
labels into an annot file but not how to make the labels from the binary m
asks.
Thank you for your attention.
2013/10/4 Bruce Fischl fischl@nmr.mgh.harvard.edu how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote: Hi, Freesurfers I am getting some interesting hippocampal surfaces with the following
steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement volumes)
3- a custom OCTAVE script to segment the concatenated volume from step
2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5 Using freeview I got a screenshot showing both hippocampi segmented
(hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6was
completed as well in 13 out 16 subjects. The only problem was during left hippocampal body smoothing in3
subjects (all the other surfaces from all 16 subjects in were smoothedwith no
problem). For these 3 surfaces, mris_smooth simply freezes, with no
error message. I must press control + C to quit the program. Duringthe
freezing I noticed high CPU use. It follows the exactly command I used and the screen output. mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody
using Gaussian curvature smoothing with norm 5.00 with 2 smooth steps
averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------------------------------ ------------------------------------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------------------------------ ------------------------------------------------------ If I supress the gaussian parameters and run just something like:
mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody I got the smoothed surface with no errors but the final quality is not
as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you cancheck
it (surfLeftHippobody). Intersting if I use a surface with all left
hippocampus (including head, problematic body and tail and) smoothing completes
without error Thank you very much for any help. MarcosThe information in this e-mail is intended only for the person to whom it
is
addressed. If you believe this e-mail was sent to you in error and the e-m
ail
contains patient information, please contact the Partners Compliance HelpL
ine at
http://www.partners.org/complianceline . If the e-mail was sent to you in
error
but does not contain patient information, please contact the sender and pr
operly
dispose of the e-mail.
ok, I just downloaded it and am trying to replicate your problem now On Tue, 8 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce Filedrop worked great and fast! Thank you for the tip. It follows the bath I run for each subject: mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippo lh.surfHippo mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoHead lh.surfHippoHead mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoTail lh.surfHippoTail mris_smooth -nw -g 5 2 -a 1 -n 1 surfRightHippo rh.surfHippo mris_smooth -nw -g 5 2 -a 1 -n 1 surfRightHippoHead rh.surfHippoHead mris_smooth -nw -g 5 2 -a 1 -n 1 surfRightHippoBody rh.surfHippoBody mris_smooth -nw -g 5 2 -a 1 -n 1 surfRightHippoTail rh.surfHippoTail
The only file with the problem I reported is surfLeftHippoBody (3rd line). If I could generate an annotation file from surfLeftHippoHead, surfLeftHippoBody and surfLeftHippoTail or the binary masks I used mri_tessellate to make these surfaces it would be a good alternative. Perhaps, even better because I would get a faster processing (smoothing 2 surfaces, surfLeftHippo and surfRightHippo, per subject instead of 8). Thank you again. Em Seg, 2013-10-07 as 20:47 -0400, Bruce Fischl escreveu:
sorry, I don't think it completed as unzip can't find the end of directory mark in the file. Try our filedrop instead:
https://gate.nmr.mgh.harvard.edu/filedrop2/
cheers Bruce
On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce I proceeded as explained at http://surfer.nmr.mgh.harvard.edu/fswiki/FtpFileExchange . Using ftp method, it started at good rate but after I transferred about 3/
4
of the file the transfer rate decreased to some bytes (not Kbytes) per second. Tried using curl -T and it seemed to complete after about 50 min b
ut
I got this message on terminal: DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head> <title>405 Method Not Allowed</title> </head><body> <h1>Method Not Allowed</h1> <p>The requested method PUT is not allowed for the URL /transfer/incoming.</p> <hr> <address>Apache/2.2.15 (CentOS) Server at surfer.nmr.mgh.harvard.edu Port 80</address> </body></html>
So I am not sure you received the file. You should find a suj46.zip corresponding to the complete folder from my subject 46. You will find the usual structure created by recon-all. At mri folder you will find all the usual volumes, the posterior* files generated by hippo-subfield processing and the files I created on my analysis. Please, could you check if you really received the file? If not, I will try again using plain ftp. Thank you again for your help. Marcos. Em Seg, 2013-10-07 as 09:21 -0400, Bruce Fischl escreveu:
hmmm, I guess you should upload the subject so that we can track it down On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce. Thank you for your fast answer. I used a script and It took less than 15 min to process all files from a
ll
subjects, except
the 3 files (surfaces) mris_smooth seems to enter a loop. When I tried t
o
smooth
surfLeftHippobody I waited for more than 4 hours. Comparing to the total
t
ime I needed to
process all files I concluded something was wrong and esed control + C t
o
interrupt the
process. I tested on 2 machines (mine is running CentOS 6 and the second
i
s a Mac, both using
Freesurfer 5.3) and got the same result. Including they seem to stop at
th
e same point. Do you
think it could need more time to finnish processing? Alternatively, as I can use mris_smooth to process the whole left hippoc
am
pus (the problem
only appears when processing the body segment), I guess I could get a si
mi
lar and perhaps
better result if I could make a label from each of the binary masks of t
he
3 segments (head,
body, tail), merge them in an annot file and finally load the whole surf
ac
e and the annot file
so I get the 3 segments discriminated by color. I found on Freesurfer Wi
ki
how to merge the
labels into an annot file but not how to make the labels from the binary
m
asks.
Thank you for your attention.
2013/10/4 Bruce Fischl fischl@nmr.mgh.harvard.edu how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote:
Hi, Freesurfers I am getting some interesting hippocampal surfaces with the
fo
llowing
steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement
v
olumes)
3- a custom OCTAVE script to segment the concatenated volume
f
rom step
2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5
Using freeview I got a screenshot showing both hippocampi se
gm
ented
(hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step
6
was
completed as well in 13 out 16 subjects.
The only problem was during left hippocampal body smoothing
in
3
subjects (all the other surfaces from all 16 subjects in were smoothe
d
with no
problem). For these 3 surfaces, mris_smooth simply freezes,
wi
th no
error message. I must press control + C to quit the program. Durin
g
the
freezing I noticed high CPU use. It follows the exactly command I used and the screen output.
mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHi
pp
oBody
using Gaussian curvature smoothing with norm 5.00 with 2 smo
ot
h steps
averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------------------------------ ------------------------------------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------------------------------ ------------------------------------------------------
If I supress the gaussian parameters and run just something
li
ke:
mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody
I got the smoothed surface with no errors but the final qual
it
y is not
as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you c
an
check
it (surfLeftHippobody). Intersting if I use a surface with all
le
ft
hippocampus (including head, problematic body and tail and) smoothing co
mp
letes
without error Thank you very much for any help. Marcos
The information in this e-mail is intended only for the person to whom i
t
is
addressed. If you believe this e-mail was sent to you in error and the e
-m
ail
contains patient information, please contact the Partners Compliance Hel
pL
ine at
http://www.partners.org/complianceline . If the e-mail was sent to you i
n
error
but does not contain patient information, please contact the sender and
pr
operly
dispose of the e-mail.
Hi Marcos
I've fixed this problem, which is a bug in one of the open source math libraries we use. If you email us the details of your hardware/software environment we'll get you a patch.
cheers Bruce On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce I proceeded as explained at http://surfer.nmr.mgh.harvard.edu/fswiki/FtpFileExchange . Using ftp method, it started at good rate but after I transferred about 3/4 of the file the transfer rate decreased to some bytes (not Kbytes) per second. Tried using curl -T and it seemed to complete after about 50 min but I got this message on terminal: DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head> <title>405 Method Not Allowed</title> </head><body> <h1>Method Not Allowed</h1> <p>The requested method PUT is not allowed for the URL /transfer/incoming.</p> <hr> <address>Apache/2.2.15 (CentOS) Server at surfer.nmr.mgh.harvard.edu Port 80</address> </body></html>
So I am not sure you received the file. You should find a suj46.zip corresponding to the complete folder from my subject 46. You will find the usual structure created by recon-all. At mri folder you will find all the usual volumes, the posterior* files generated by hippo-subfield processing and the files I created on my analysis. Please, could you check if you really received the file? If not, I will try again using plain ftp. Thank you again for your help. Marcos. Em Seg, 2013-10-07 às 09:21 -0400, Bruce Fischl escreveu:
hmmm, I guess you should upload the subject so that we can track it down On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce. Thank you for your fast answer. I used a script and It took less than 15 min to process all files from all subjects, except the 3 files (surfaces) mris_smooth seems to enter a loop. When I tried to smooth surfLeftHippobody I waited for more than 4 hours. Comparing to the total time I needed to process all files I concluded something was wrong and esed control + C to interrupt the process. I tested on 2 machines (mine is running CentOS 6 and the second is a Mac, both using Freesurfer 5.3) and got the same result. Including they seem to stop at the same point. Do you think it could need more time to finnish processing? Alternatively, as I can use mris_smooth to process the whole left hippocampus (the problem only appears when processing the body segment), I guess I could get a similar and perhaps better result if I could make a label from each of the binary masks of the 3 segments (head, body, tail), merge them in an annot file and finally load the whole surface and the annot file so I get the 3 segments discriminated by color. I found on Freesurfer Wiki how to merge the labels into an annot file but not how to make the labels from the binary masks. Thank you for your attention.
2013/10/4 Bruce Fischl fischl@nmr.mgh.harvard.edu how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote:
Hi, Freesurfers I am getting some interesting hippocampal surfaces with the following steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement volumes) 3- a custom OCTAVE script to segment the concatenated volume from step 2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5
Using freeview I got a screenshot showing both hippocampi segmented (hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6 was completed as well in 13 out 16 subjects.
The only problem was during left hippocampal body smoothing in 3 subjects (all the other surfaces from all 16 subjects in were smoothed with no problem). For these 3 surfaces, mris_smooth simply freezes, with no error message. I must press control + C to quit the program. During the freezing I noticed high CPU use. It follows the exactly command I used and the screen output.
mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody using Gaussian curvature smoothing with norm 5.00 with 2 smooth steps averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------------------------------ ------------------------------------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------------------------------ ------------------------------------------------------
If I supress the gaussian parameters and run just something like: mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody
I got the smoothed surface with no errors but the final quality is not as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you can check it (surfLeftHippobody). Intersting if I use a surface with all left hippocampus (including head, problematic body and tail and) smoothing completes without error Thank you very much for any help. Marcos
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Wow! That was fast. I am running FreeSurfer 5.3 on top of CentOS6 (64-bit). My hardware is a PC powered by Intel® Core™ i5 CPU M 520 @ 2.40GHz × 4 Thank you for all help. Cheers Marcos. Em Ter, 2013-10-08 às 09:00 -0400, Bruce Fischl escreveu:
Hi Marcos
I've fixed this problem, which is a bug in one of the open source math libraries we use. If you email us the details of your hardware/software environment we'll get you a patch.
cheers Bruce On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce I proceeded as explained at http://surfer.nmr.mgh.harvard.edu/fswiki/FtpFileExchange . Using ftp method, it started at good rate but after I transferred about 3/4 of the file the transfer rate decreased to some bytes (not Kbytes) per second. Tried using curl -T and it seemed to complete after about 50 min but I got this message on terminal: DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head> <title>405 Method Not Allowed</title> </head><body> <h1>Method Not Allowed</h1> <p>The requested method PUT is not allowed for the URL /transfer/incoming.</p> <hr> <address>Apache/2.2.15 (CentOS) Server at surfer.nmr.mgh.harvard.edu Port 80</address> </body></html>
So I am not sure you received the file. You should find a suj46.zip corresponding to the complete folder from my subject 46. You will find the usual structure created by recon-all. At mri folder you will find all the usual volumes, the posterior* files generated by hippo-subfield processing and the files I created on my analysis. Please, could you check if you really received the file? If not, I will try again using plain ftp. Thank you again for your help. Marcos. Em Seg, 2013-10-07 às 09:21 -0400, Bruce Fischl escreveu:
hmmm, I guess you should upload the subject so that we can track it down On Mon, 7 Oct 2013, Marcos Martins da Silva wrote:
Hi, Bruce. Thank you for your fast answer. I used a script and It took less than 15 min to process all files from all subjects, except the 3 files (surfaces) mris_smooth seems to enter a loop. When I tried to smooth surfLeftHippobody I waited for more than 4 hours. Comparing to the total time I needed to process all files I concluded something was wrong and esed control + C to interrupt the process. I tested on 2 machines (mine is running CentOS 6 and the second is a Mac, both using Freesurfer 5.3) and got the same result. Including they seem to stop at the same point. Do you think it could need more time to finnish processing? Alternatively, as I can use mris_smooth to process the whole left hippocampus (the problem only appears when processing the body segment), I guess I could get a similar and perhaps better result if I could make a label from each of the binary masks of the 3 segments (head, body, tail), merge them in an annot file and finally load the whole surface and the annot file so I get the 3 segments discriminated by color. I found on Freesurfer Wiki how to merge the labels into an annot file but not how to make the labels from the binary masks. Thank you for your attention.
2013/10/4 Bruce Fischl fischl@nmr.mgh.harvard.edu how long have you waited?
On Fri, 4 Oct 2013, Marcos Martins da Silva wrote: Hi, Freesurfers I am getting some interesting hippocampal surfaces with the following steps: 1- recon-all -all -hippo-subfields 2- mri-concat (only CA1, CA23, CA4DG and the tail complement volumes) 3- a custom OCTAVE script to segment the concatenated volume from step 2 into Head, Body and Tail 4- mri_binarize Head, Body and Tail volumes from step 3 5- mri_tessellate the binary masks from step 4 6- mri_smooth the surfaces from step 5 Using freeview I got a screenshot showing both hippocampi segmented (hipposurf3d.png is attached). This approach worked well for 16 subjects up to step 5. Step 6 was completed as well in 13 out 16 subjects. The only problem was during left hippocampal body smoothing in 3 subjects (all the other surfaces from all 16 subjects in were smoothed with no problem). For these 3 surfaces, mris_smooth simply freezes, with no error message. I must press control + C to quit the program. During the freezing I noticed high CPU use. It follows the exactly command I used and the screen output. mris_smooth -nw -g 5 2 -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody using Gaussian curvature smoothing with norm 5.00 with 2 smooth steps averaging curvature for 1 iterations smoothing for 1 iterations smoothing surface tessellation for 1 iterations... ------------------------------------------------------ ------------------------------------------------------ ------------------ pass 1 (num=2853) ------------------ ------------------------------------------------------ ------------------------------------------------------ If I supress the gaussian parameters and run just something like: mris_smooth -nw -a 1 -n 1 surfLeftHippoBody lh.surfHippoBody or even mris_smooth -nw surfLeftHippoBody lh.surfHippoBody I got the smoothed surface with no errors but the final quality is not as good as when I use the gaussian parameters. I am attaching one of the problematic surface files so you can check it (surfLeftHippobody). Intersting if I use a surface with all left hippocampus (including head, problematic body and tail and) smoothing completes without error Thank you very much for any help. MarcosThe information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
Marcos Martins da Silva