External Email - Use Caution
Dear FreeSurfer experts,
I am running mri_deface for all relevant volume files of more than 700 subjects (~3500 MGZ files in total). For roughly 10% of the files, the command fails, i.e., it stops with non-zero exit code and does not produce a result file.
The relevant part of the mri_deface output is:
after smoothing, mri peak at 0, scaling input intensities by inf GCAhistoScaleImageIntensities: could not find wm peak
The full command line, an example file and the full output can be found in the footer of this email.
Is there anything I can try to fix this?
Some more information/observations:
I noticed that for the subjects with errors, there are almost *always* the same 3 files affected: * mri/orig.mgz * mri/rawavg.mgz * mri/orig/001.mgz
The file mri/T1.mgz is never affected. The file mri/orig_nu.mgz is only affected 2 times in total, and in these cases the image quality is extremely bad, so they can be ignored.
Therefore, I suspect that the problems with the orig/rawavg/001 files are related to intensity inhomogeneity, which may lead to issues with the registration to the skull/face templates.
Tim
-- Dr. Tim Schäfer Postdoc Computational Neuroimaging Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy University Hospital Frankfurt, Goethe University Frankfurt am Main, Germany
* Demo file: rawavg.mgz uploaded via FTP to /transfer/incoming/rawavg.mgz on surfer.nmr.mgh.harvard.edu * System info: FreeSurfer v6 stable, MacOS * Full mri_deface command and output for the demo file:
$ SKULL_TEMPLATE="${FREESURFER_HOME}/average/talairach_mixed_with_skull.gca" $ FACE_TEMPLATE="${FREESURFER_HOME}/average/face.gca" $ mri_deface /path/to/subject/mri/rawavg.mgz "${SKULL_TEMPLATE}" "${FACE_TEMPLATE}" ~/defaced.mgz logging results to defaced.log reading '/Applications/freesurfer/average/talairach_mixed_with_skull.gca'... reading '/path/to/subject/mri/rawavg.mgz'... changing type of input volume to 8 bits/voxel... MRIchangeType: Building histogram bounding unknown intensity as < 10.3 or > 1461.6 total sample mean = 111.8 (0 zeros) spacing=8, using 2055 sample points, tol=1.00e-03... resetting wm mean[0]: 143 --> 144 resetting gm mean[0]: 89 --> 89 input volume #1 is the most T1-like using real data threshold=6.0 skull bounding box = (40, 39, 17) --> (213, 255, 225) using (98, 111, 121) as brain centroid... mean wm in atlas = 144, using box (77,84,95) --> (119, 137,146) to find MRI wm before smoothing, mri peak at 89 robust fit to distribution - 89 +- 6.7 distribution too broad for accurate scaling - disabling WARNING2: gca.c::GCAhistoScaleImageIntensities: h_mri->nbins=109, mri_peak=141 after smoothing, mri peak at 0, scaling input intensities by inf GCAhistoScaleImageIntensities: could not find wm peak
Hi Tim
if you upload the input files to mri_deface and the full command line I'll take a look. Is this the dev version you are using?
cheers Bruce On Mon, 3 Feb 2020, Tim Schäfer wrote:
External Email - Use CautionDear FreeSurfer experts,
I am running mri_deface for all relevant volume files of more than 700 subjects (~3500 MGZ files in total). For roughly 10% of the files, the command fails, i.e., it stops with non-zero exit code and does not produce a result file.
The relevant part of the mri_deface output is:
after smoothing, mri peak at 0, scaling input intensities by inf GCAhistoScaleImageIntensities: could not find wm peak
The full command line, an example file and the full output can be found in the footer of this email.
Is there anything I can try to fix this?
Some more information/observations:
I noticed that for the subjects with errors, there are almost *always* the same 3 files affected:
- mri/orig.mgz
- mri/rawavg.mgz
- mri/orig/001.mgz
The file mri/T1.mgz is never affected. The file mri/orig_nu.mgz is only affected 2 times in total, and in these cases the image quality is extremely bad, so they can be ignored.
Therefore, I suspect that the problems with the orig/rawavg/001 files are related to intensity inhomogeneity, which may lead to issues with the registration to the skull/face templates.
Tim
-- Dr. Tim Schäfer Postdoc Computational Neuroimaging Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy University Hospital Frankfurt, Goethe University Frankfurt am Main, Germany
- Demo file: rawavg.mgz uploaded via FTP to /transfer/incoming/rawavg.mgz on surfer.nmr.mgh.harvard.edu
- System info: FreeSurfer v6 stable, MacOS
- Full mri_deface command and output for the demo file:
$ SKULL_TEMPLATE="${FREESURFER_HOME}/average/talairach_mixed_with_skull.gca" $ FACE_TEMPLATE="${FREESURFER_HOME}/average/face.gca" $ mri_deface /path/to/subject/mri/rawavg.mgz "${SKULL_TEMPLATE}" "${FACE_TEMPLATE}" ~/defaced.mgz logging results to defaced.log reading '/Applications/freesurfer/average/talairach_mixed_with_skull.gca'... reading '/path/to/subject/mri/rawavg.mgz'... changing type of input volume to 8 bits/voxel... MRIchangeType: Building histogram bounding unknown intensity as < 10.3 or > 1461.6 total sample mean = 111.8 (0 zeros) spacing=8, using 2055 sample points, tol=1.00e-03... resetting wm mean[0]: 143 --> 144 resetting gm mean[0]: 89 --> 89 input volume #1 is the most T1-like using real data threshold=6.0 skull bounding box = (40, 39, 17) --> (213, 255, 225) using (98, 111, 121) as brain centroid... mean wm in atlas = 144, using box (77,84,95) --> (119, 137,146) to find MRI wm before smoothing, mri peak at 89 robust fit to distribution - 89 +- 6.7 distribution too broad for accurate scaling - disabling WARNING2: gca.c::GCAhistoScaleImageIntensities: h_mri->nbins=109, mri_peak=141 after smoothing, mri peak at 0, scaling input intensities by inf GCAhistoScaleImageIntensities: could not find wm peak
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
External Email - Use Caution
Hi Bruce,
I have uploaded an input file via FTP to /transfer/incoming/rawavg.mgz on ftp://surfer.nmr.mgh.harvard.edu.
I am using FreeSurfer v6 stable under MacOS.
The full command line (bash shell) is:
cd path/to/rawavg.mgz mri_deface rawavg.mgz $FREESURFER_HOME/average/talairach_mixed_with_skull.gca $FREESURFER_HOME/average/face.gca defaced.mgz
Thanks for your time and all the best,
Tim
On February 3, 2020 at 4:10 PM Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Tim
if you upload the input files to mri_deface and the full command line I'll take a look. Is this the dev version you are using?
cheers Bruce On Mon, 3 Feb 2020, Tim Schäfer wrote:
External Email - Use CautionDear FreeSurfer experts,
I am running mri_deface for all relevant volume files of more than 700 subjects (~3500 MGZ files in total). For roughly 10% of the files, the command fails, i.e., it stops with non-zero exit code and does not produce a result file.
The relevant part of the mri_deface output is:
after smoothing, mri peak at 0, scaling input intensities by inf GCAhistoScaleImageIntensities: could not find wm peak
The full command line, an example file and the full output can be found in the footer of this email.
Is there anything I can try to fix this?
Some more information/observations:
I noticed that for the subjects with errors, there are almost *always* the same 3 files affected:
- mri/orig.mgz
- mri/rawavg.mgz
- mri/orig/001.mgz
The file mri/T1.mgz is never affected. The file mri/orig_nu.mgz is only affected 2 times in total, and in these cases the image quality is extremely bad, so they can be ignored.
Therefore, I suspect that the problems with the orig/rawavg/001 files are related to intensity inhomogeneity, which may lead to issues with the registration to the skull/face templates.
Tim
-- Dr. Tim Schäfer Postdoc Computational Neuroimaging Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy University Hospital Frankfurt, Goethe University Frankfurt am Main, Germany
- Demo file: rawavg.mgz uploaded via FTP to /transfer/incoming/rawavg.mgz on surfer.nmr.mgh.harvard.edu
- System info: FreeSurfer v6 stable, MacOS
- Full mri_deface command and output for the demo file:
$ SKULL_TEMPLATE="${FREESURFER_HOME}/average/talairach_mixed_with_skull.gca" $ FACE_TEMPLATE="${FREESURFER_HOME}/average/face.gca" $ mri_deface /path/to/subject/mri/rawavg.mgz "${SKULL_TEMPLATE}" "${FACE_TEMPLATE}" ~/defaced.mgz logging results to defaced.log reading '/Applications/freesurfer/average/talairach_mixed_with_skull.gca'... reading '/path/to/subject/mri/rawavg.mgz'... changing type of input volume to 8 bits/voxel... MRIchangeType: Building histogram bounding unknown intensity as < 10.3 or > 1461.6 total sample mean = 111.8 (0 zeros) spacing=8, using 2055 sample points, tol=1.00e-03... resetting wm mean[0]: 143 --> 144 resetting gm mean[0]: 89 --> 89 input volume #1 is the most T1-like using real data threshold=6.0 skull bounding box = (40, 39, 17) --> (213, 255, 225) using (98, 111, 121) as brain centroid... mean wm in atlas = 144, using box (77,84,95) --> (119, 137,146) to find MRI wm before smoothing, mri peak at 89 robust fit to distribution - 89 +- 6.7 distribution too broad for accurate scaling - disabling WARNING2: gca.c::GCAhistoScaleImageIntensities: h_mri->nbins=109, mri_peak=141 after smoothing, mri peak at 0, scaling input intensities by inf GCAhistoScaleImageIntensities: could not find wm peak
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 -
Tim Schäfer