External Email - Use Caution
Hi there,
I wonder if its somehow possible to feed FreeSurfer's recon-all with pre-existing segmented maps (based on prior segmentation procedure e.g. SPM or ANTs) to "improve" FreeSurfer results.
For some images I am working with the FreeSurfer's WM segmentation output is not optimal so I am looking for alternatives to improve the final results on such specific cases without requiring any manual intervention.
Any comments, examples or shared reflections on this regard are kindly welcome.
Thanks in advance, Jordi Huguet
Hi Jordi
yes, this should be possible. You might need to run recon-all in stages though. Although if you set the wm.mgz values from SPM to 255 and 1 (not 0) recon-all should detect them as "edits" and retain them
cheers Bruce
On Tue, 26 Nov 2019, Jordi Huguet wrote:
External Email - Use Caution
Hi there, I wonder if its somehow possible to feed FreeSurfer's recon-all with pre-existing segmented maps (based on prior segmentation procedure e.g. SPM or ANTs) to "improve" FreeSurfer results.
For some images I am working with the FreeSurfer's WM segmentation output is not optimal so I am looking for alternatives to improve the final results on such specific cases without requiring any manual intervention.
Any comments, examples or shared reflections on this regard are kindly welcome.
Thanks in advance, Jordi Huguet
External Email - Use Caution
Hi Bruce,
Many thanks for your reply.
I understand from your answer that it could be done as follows (I am not an expert in FreeSurfer directives):
a) run full FreeSurfer pipeline recon-all -s FOO b) insert pre-existing WM segmentation as wm.mgz (modifying the intensity values) c) re-run stages 15-23 and 24-31 (recon-all shall use my external segmentation as WM "edits") recon-all -s FOO -autorecon2-wm -autorecon3
Is it possible however to just include the WM segmentation without executing a full early run of recon-all?
Thanks, Jordi
On Tue, Nov 26, 2019 at 3:08 PM Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Jordi
yes, this should be possible. You might need to run recon-all in stages though. Although if you set the wm.mgz values from SPM to 255 and 1 (not 0) recon-all should detect them as "edits" and retain them
cheers Bruce
On Tue, 26 Nov 2019, Jordi Huguet wrote:
External Email - Use CautionHi there, I wonder if its somehow possible to feed FreeSurfer's recon-all with
pre-existing segmented maps
(based on prior segmentation procedure e.g. SPM or ANTs) to
"improve" FreeSurfer results.
For some images I am working with the FreeSurfer's WM segmentation
output is not optimal so I am
looking for alternatives to improve the final results on such specific
cases without requiring any
manual intervention.
Any comments, examples or shared reflections on this regard are kindly
welcome.
Thanks in advance, Jordi Huguet
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
It might work if you mkdir -p FOO/mri cp yourwm.mgz FOO/mri/wm.mgz cp inputdata.mgz FOO/mri/orig/001.mgz then run recon-all -s FOO -all
On 11/27/2019 12:39 PM, Jordi Huguet wrote:
External Email - Use Caution
Hi Bruce,
Many thanks for your reply.
I understand from your answer that it could be done as follows (I am not an expert in FreeSurfer directives):
a) run full FreeSurfer pipeline recon-all -s FOO b) insert pre-existing WM segmentation as wm.mgz (modifying the intensity values) c) re-run stages 15-23 and 24-31 (recon-all shall use my external segmentation as WM "edits") recon-all -s FOO -autorecon2-wm -autorecon3
Is it possible however to just include the WM segmentation without executing a full early run of recon-all?
Thanks, Jordi
On Tue, Nov 26, 2019 at 3:08 PM Bruce Fischl <fischl@nmr.mgh.harvard.edumailto:fischl@nmr.mgh.harvard.edu> wrote: Hi Jordi
yes, this should be possible. You might need to run recon-all in stages though. Although if you set the wm.mgz values from SPM to 255 and 1 (not 0) recon-all should detect them as "edits" and retain them
cheers Bruce
On Tue, 26 Nov 2019, Jordi Huguet wrote:
External Email - Use CautionHi there, I wonder if its somehow possible to feed FreeSurfer's recon-all with pre-existing segmented maps (based on prior segmentation procedure e.g. SPM or ANTs) to "improve" FreeSurfer results.
For some images I am working with the FreeSurfer's WM segmentation output is not optimal so I am looking for alternatives to improve the final results on such specific cases without requiring any manual intervention.
Any comments, examples or shared reflections on this regard are kindly welcome.
Thanks in advance, Jordi Huguet
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
External Email - Use Caution
Great, just what I was looking for!
Thanks Douglas, Jordi
On Wed, Nov 27, 2019, 19:47 Greve, Douglas N.,Ph.D. DGREVE@mgh.harvard.edu wrote:
It might work if you mkdir -p FOO/mri cp yourwm.mgz FOO/mri/wm.mgz cp inputdata.mgz FOO/mri/orig/001.mgz then run recon-all -s FOO -all
On 11/27/2019 12:39 PM, Jordi Huguet wrote:
External Email - Use CautionHi Bruce,
Many thanks for your reply.
I understand from your answer that it could be done as follows (I am not an expert in FreeSurfer directives):
a) run full FreeSurfer pipeline recon-all -s FOO b) insert pre-existing WM segmentation as wm.mgz (modifying the intensity values) c) re-run stages 15-23 and 24-31 (recon-all shall use my external segmentation as WM "edits") recon-all -s FOO -autorecon2-wm -autorecon3
Is it possible however to just include the WM segmentation without executing a full early run of recon-all?
Thanks, Jordi
On Tue, Nov 26, 2019 at 3:08 PM Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Jordi
yes, this should be possible. You might need to run recon-all in stages though. Although if you set the wm.mgz values from SPM to 255 and 1 (not 0) recon-all should detect them as "edits" and retain them
cheers Bruce
On Tue, 26 Nov 2019, Jordi Huguet wrote:
External Email - Use CautionHi there, I wonder if its somehow possible to feed FreeSurfer's recon-all with
pre-existing segmented maps
(based on prior segmentation procedure e.g. SPM or ANTs) to
"improve" FreeSurfer results.
For some images I am working with the FreeSurfer's WM segmentation
output is not optimal so I am
looking for alternatives to improve the final results on such specific
cases without requiring any
manual intervention.
Any comments, examples or shared reflections on this regard are kindly
welcome.
Thanks in advance, Jordi Huguet
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
External Email - Use Caution
Hi Douglas,
I have followed the suggested approach, however I get an error related to the dimensions of the WM volume (see log trace below).
...
Found wm edits: 23448857 deletes, 783317 fills
cp wm.mgz wm.seg.mgz
mri_segment -keep -mprage brain.mgz wm.seg.mgz
preserving editing changes in output volume... doing initial intensity segmentation... using local statistics to label ambiguous voxels... computing class statistics for intensity windows... WM (104.0): 105.4 +- 6.0 [79.0 --> 125.0] GM (69.0) : 67.0 +- 10.5 [30.0 --> 95.0] setting bottom of white matter range to 77.5 setting top of gray matter range to 88.0 doing initial intensity segmentation... using local statistics to label ambiguous voxels... using local geometry to label remaining ambiguous voxels...
reclassifying voxels using Gaussian border classifier...
removing voxels with positive offset direction... smoothing T1 volume with sigma = 0.250 removing 1-dimensional structures... thickening thin strands.... 20 segments, 7509 filled 8232 bright non-wm voxels segmented. 8121 diagonally connected voxels added... white matter segmentation took 2.7 minutes writing output to wm.seg.mgz... *ERROR: mri_segment-MRIcheckVolDims: volume1 width=240 != volume2 width=320.*
My WM segmentations (as well as the T1w images) are 320 rows x 320 columns in 240 frames. Seems like FreeSurfer expects them to be cubic, correct? Would that fix the problem if I add 80 frames of zeroes (or ones) to my volumes? Any other ideas of what I might be doing wrong?
Regards, Jordi
On Thu, Nov 28, 2019 at 8:47 AM Jordi Huguet jhuguetn@gmail.com wrote:
Great, just what I was looking for!
Thanks Douglas, Jordi
On Wed, Nov 27, 2019, 19:47 Greve, Douglas N.,Ph.D. < DGREVE@mgh.harvard.edu> wrote:
It might work if you mkdir -p FOO/mri cp yourwm.mgz FOO/mri/wm.mgz cp inputdata.mgz FOO/mri/orig/001.mgz then run recon-all -s FOO -all
On 11/27/2019 12:39 PM, Jordi Huguet wrote:
External Email - Use CautionHi Bruce,
Many thanks for your reply.
I understand from your answer that it could be done as follows (I am not an expert in FreeSurfer directives):
a) run full FreeSurfer pipeline recon-all -s FOO b) insert pre-existing WM segmentation as wm.mgz (modifying the intensity values) c) re-run stages 15-23 and 24-31 (recon-all shall use my external segmentation as WM "edits") recon-all -s FOO -autorecon2-wm -autorecon3
Is it possible however to just include the WM segmentation without executing a full early run of recon-all?
Thanks, Jordi
On Tue, Nov 26, 2019 at 3:08 PM Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Jordi
yes, this should be possible. You might need to run recon-all in stages though. Although if you set the wm.mgz values from SPM to 255 and 1 (not 0) recon-all should detect them as "edits" and retain them
cheers Bruce
On Tue, 26 Nov 2019, Jordi Huguet wrote:
External Email - Use CautionHi there, I wonder if its somehow possible to feed FreeSurfer's recon-all with
pre-existing segmented maps
(based on prior segmentation procedure e.g. SPM or ANTs) to
"improve" FreeSurfer results.
For some images I am working with the FreeSurfer's WM segmentation
output is not optimal so I am
looking for alternatives to improve the final results on such specific
cases without requiring any
manual intervention.
Any comments, examples or shared reflections on this regard are kindly
welcome.
Thanks in advance, Jordi Huguet
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
If your WM volume is not conformed, you will need to conform it with mri_convert yourwm.mgz --conform wm.mgz
On 12/5/19 4:24 AM, Jordi Huguet wrote:
External Email - Use Caution
Hi Douglas,
I have followed the suggested approach, however I get an error related to the dimensions of the WM volume (see log trace below).
... Found wm edits: 23448857 deletes, 783317 fills cp wm.mgz wm.seg.mgz mri_segment -keep -mprage brain.mgz wm.seg.mgz preserving editing changes in output volume... doing initial intensity segmentation... using local statistics to label ambiguous voxels... computing class statistics for intensity windows... WM (104.0): 105.4 +- 6.0 [79.0 --> 125.0] GM (69.0) : 67.0 +- 10.5 [30.0 --> 95.0] setting bottom of white matter range to 77.5 setting top of gray matter range to 88.0 doing initial intensity segmentation... using local statistics to label ambiguous voxels... using local geometry to label remaining ambiguous voxels... reclassifying voxels using Gaussian border classifier... removing voxels with positive offset direction... smoothing T1 volume with sigma = 0.250 removing 1-dimensional structures... thickening thin strands.... 20 segments, 7509 filled 8232 bright non-wm voxels segmented. 8121 diagonally connected voxels added... white matter segmentation took 2.7 minutes writing output to wm.seg.mgz... *ERROR: mri_segment-MRIcheckVolDims: volume1 width=240 != volume2 width=320.*My WM segmentations (as well as the T1w images) are 320 rows x 320 columns in 240 frames. Seems like FreeSurfer expects them to be cubic, correct? Would that fix the problem if I add 80 frames of zeroes (or ones) to my volumes? Any other ideas of what I might be doing wrong?
Regards, Jordi
On Thu, Nov 28, 2019 at 8:47 AM Jordi Huguet <jhuguetn@gmail.com mailto:jhuguetn@gmail.com> wrote:
Great, just what I was looking for! Thanks Douglas, Jordi On Wed, Nov 27, 2019, 19:47 Greve, Douglas N.,Ph.D. <DGREVE@mgh.harvard.edu <mailto:DGREVE@mgh.harvard.edu>> wrote: It might work if you mkdir -p FOO/mri cp yourwm.mgz FOO/mri/wm.mgz cp inputdata.mgz FOO/mri/orig/001.mgz then run recon-all -s FOO -all On 11/27/2019 12:39 PM, Jordi Huguet wrote:External Email - Use Caution Hi Bruce, Many thanks for your reply. I understand from your answer that it could be done as follows (I am not an expert in FreeSurfer directives): a) run full FreeSurfer pipeline recon-all -s FOO b) insert pre-existing WM segmentation as wm.mgz (modifying the intensity values) c) re-run stages 15-23 and 24-31 (recon-all shall use my external segmentation as WM "edits") recon-all -s FOO -autorecon2-wm -autorecon3 Is it possible however to just include the WM segmentation without executing a full early run of recon-all? Thanks, Jordi On Tue, Nov 26, 2019 at 3:08 PM Bruce Fischl <fischl@nmr.mgh.harvard.edu <mailto:fischl@nmr.mgh.harvard.edu>> wrote: Hi Jordi yes, this should be possible. You might need to run recon-all in stages though. Although if you set the wm.mgz values from SPM to 255 and 1 (not 0) recon-all should detect them as "edits" and retain them cheers Bruce On Tue, 26 Nov 2019, Jordi Huguet wrote: > > External Email - Use Caution > > Hi there, > I wonder if its somehow possible to feed FreeSurfer's recon-all with pre-existing segmented maps > (based on prior segmentation procedure e.g. SPM or ANTs) to "improve" FreeSurfer results. > > For some images I am working with the FreeSurfer's WM segmentation output is not optimal so I am > looking for alternatives to improve the final results on such specific cases without requiring any > manual intervention. > > Any comments, examples or shared reflections on this regard are kindly welcome. > > Thanks in advance, > Jordi Huguet > >_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
External Email - Use Caution
That's what I needed, thanks! It all worked well now.
Actually using '--conform_min' preserved existing image dimensions as desired, otherwise image got reshaped to 256^3 raising the aforementioned dimensionality error.
Jordi
On Thu, Dec 5, 2019, 18:36 Greve, Douglas N.,Ph.D. DGREVE@mgh.harvard.edu wrote:
If your WM volume is not conformed, you will need to conform it with mri_convert yourwm.mgz --conform wm.mgz
On 12/5/19 4:24 AM, Jordi Huguet wrote:
External Email - Use CautionHi Douglas,
I have followed the suggested approach, however I get an error related to the dimensions of the WM volume (see log trace below).
... Found wm edits: 23448857 deletes, 783317 fills cp wm.mgz wm.seg.mgz mri_segment -keep -mprage brain.mgz wm.seg.mgz preserving editing changes in output volume... doing initial intensity segmentation... using local statistics to label ambiguous voxels... computing class statistics for intensity windows... WM (104.0): 105.4 +- 6.0 [79.0 --> 125.0] GM (69.0) : 67.0 +- 10.5 [30.0 --> 95.0] setting bottom of white matter range to 77.5 setting top of gray matter range to 88.0 doing initial intensity segmentation... using local statistics to label ambiguous voxels... using local geometry to label remaining ambiguous voxels... reclassifying voxels using Gaussian border classifier... removing voxels with positive offset direction... smoothing T1 volume with sigma = 0.250 removing 1-dimensional structures... thickening thin strands.... 20 segments, 7509 filled 8232 bright non-wm voxels segmented. 8121 diagonally connected voxels added... white matter segmentation took 2.7 minutes writing output to wm.seg.mgz... *ERROR: mri_segment-MRIcheckVolDims: volume1 width=240 != volume2 width=320.*My WM segmentations (as well as the T1w images) are 320 rows x 320 columns in 240 frames. Seems like FreeSurfer expects them to be cubic, correct? Would that fix the problem if I add 80 frames of zeroes (or ones) to my volumes? Any other ideas of what I might be doing wrong?
Regards, Jordi
On Thu, Nov 28, 2019 at 8:47 AM Jordi Huguet <jhuguetn@gmail.com mailto:jhuguetn@gmail.com> wrote:
Great, just what I was looking for! Thanks Douglas, Jordi On Wed, Nov 27, 2019, 19:47 Greve, Douglas N.,Ph.D. <DGREVE@mgh.harvard.edu <mailto:DGREVE@mgh.harvard.edu>> wrote: It might work if you mkdir -p FOO/mri cp yourwm.mgz FOO/mri/wm.mgz cp inputdata.mgz FOO/mri/orig/001.mgz then run recon-all -s FOO -all On 11/27/2019 12:39 PM, Jordi Huguet wrote:External Email - Use Caution Hi Bruce, Many thanks for your reply. I understand from your answer that it could be done as follows (I am not an expert in FreeSurfer directives): a) run full FreeSurfer pipeline recon-all -s FOO b) insert pre-existing WM segmentation as wm.mgz (modifying the intensity values) c) re-run stages 15-23 and 24-31 (recon-all shall use my external segmentation as WM "edits") recon-all -s FOO -autorecon2-wm -autorecon3 Is it possible however to just include the WM segmentation without executing a full early run of recon-all? Thanks, Jordi On Tue, Nov 26, 2019 at 3:08 PM Bruce Fischl <fischl@nmr.mgh.harvard.edu <mailto:fischl@nmr.mgh.harvard.edu>> wrote: Hi Jordi yes, this should be possible. You might need to run recon-all in stages though. Although if you set the wm.mgz values from SPM to 255 and 1 (not 0) recon-all should detect them as "edits" and retain them cheers Bruce On Tue, 26 Nov 2019, Jordi Huguet wrote: > > External Email - Use Caution > > Hi there, > I wonder if its somehow possible to feed FreeSurfer's recon-all with pre-existing segmented maps > (based on prior segmentation procedure e.g. SPM or ANTs) to "improve" FreeSurfer results. > > For some images I am working with the FreeSurfer's WM segmentation output is not optimal so I am > looking for alternatives to improve the final results on such specific cases without requiring any > manual intervention. > > Any comments, examples or shared reflections on this regard are kindly welcome. > > Thanks in advance, > Jordi Huguet > >_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
Greve, Douglas N.,Ph.D. -
Jordi Huguet