External Email - Use Caution
Hi Sally, It is indeed the linear registration used to initialize the atlas deformation that is going crazy. Do you think you could upload the subject, so we can take a look? Cheers, /Eugenio
-- Juan Eugenio Iglesias ERC Senior Research Fellow Centre for Medical Image Computing (CMIC) University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
From: freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Sally Grace sgrace@swin.edu.au Reply-To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Date: Monday, 8 October 2018 at 06:21 To: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Hippocampal/Amygdala subfield segmentation error - Freesurfer devel-20180612
External Email - Use Caution Dear experts, I am running the hippocampal/amygdala segmentation on T1's preprocessed in -reconall in FS 5.3 using FS devel-20180612 using the command segmentHA_T1.sh bert [SUBJECTS_DIR] This has worked successfully for roughly 1000 participants but for one participant I keep getting the error below, despite several troubleshooting attempts. The error appears to be "Fitting mesh to synthetic image resulted in no deformation" - I have had a look at the hippocampus masks and they appear to be registered incorrectly, despite the grey matter volume output from -reconall having no errors. Do you have any idea what I can do to troubleshoot or fix this? Thanks in advance, Sally.
Error Log:
------------------------------ USER valentil HOST m3a007 PROCESSID 29440 PROCESSOR x86_64 OS Linux Linux m3a007 3.10.0-514.26.1.el7.x86_64 #1 SMP Thu Jun 29 16:05:25 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux ------------------------------
#-------------------------------------------- #@# Hippocampal Subfields processing (T1) left Tue Sep 25 10:18:56 AEST 2018 ------------------------------------------ Setting up environment variables --- LD_LIBRARY_PATH is .:/usr/local/freesurfer/devel-20180612/MCRv84//runtime/glnxa64:/usr/local/freesurfer/devel-20180612/MCRv84//bin/glnxa64:/usr/local/freesurfer/devel-20180612/MCRv84//sys/os/glnxa64:/usr/local/freesurfer/devel-20180612/lib/tcltktixblt:/usr/local/freesurfer/devel-20180612/lib/qt/lib:/usr/local/freesurfer/devel-20180612/lib/petsc/lib:/usr/local/freesurfer/devel-20180612/lib/KWWidgets/lib/KWWidgets:/usr/local/freesurfer/devel-20180612/lib/cuda/lib64:/usr/local/freesurfer/devel-20180612/lib/cuda/lib:/usr/local/freesurfer/devel-20180612/lib/bem:/usr/local/freesurfer/devel-20180612/lib:/usr/local/freesurfer/devel-20180612/lib/vtk/lib/vtk-5.6:/usr/local/fsl/5.0.11/fsl/lib:/usr/local/libjpeg-turbo/1.4.2/lib64:/usr/local/cuda/7.5/extras/CUPTI/lib64:/usr/local/cuda/7.5/lib64:/usr/local/cuda/7.5/lib:/usr/local/cuda/7.5/lib64/stubs:/opt/munge-0.5.11/lib:/opt/slurm-17.11.4/lib:/opt/slurm-17.11.4/lib/slurm:/usr/local/tigervnc/1.8.0/lib64:/usr/local/tigervnc/1.8.0/lib:/opt/munge-0.5.11/lib:/opt/slurm-17.11.4/lib:/opt/slurm-17.11.4/lib/slurm::/usr/local/freesurfer/devel-20180612/MCRv84//sys/opengl/lib/glnxa64:/usr/local/freesurfer/devel-20180612/lib/tcltktixblt:/usr/local/freesurfer/devel-20180612/lib/qt/lib:/usr/local/freesurfer/devel-20180612/lib/petsc/lib:/usr/local/freesurfer/devel-20180612/lib/KWWidgets/lib/KWWidgets:/usr/local/freesurfer/devel-20180612/lib/cuda/lib64:/usr/local/freesurfer/devel-20180612/lib/cuda/lib:/usr/local/freesurfer/devel-20180612/lib/bem:/usr/local/freesurfer/devel-20180612/lib:/usr/local/freesurfer/devel-20180612/lib/vtk/lib/vtk-5.6:/usr/local/fsl/5.0.11/fsl/lib:/usr/local/libjpeg-turbo/1.4.2/lib64:/usr/local/cuda/7.5/extras/CUPTI/lib64:/usr/local/cuda/7.5/lib64:/usr/local/cuda/7.5/lib:/usr/local/cuda/7.5/lib64/stubs:/opt/munge-0.5.11/lib:/opt/slurm-17.11.4/lib:/opt/slurm-17.11.4/lib/slurm:/usr/local/tigervnc/1.8.0/lib64:/usr/local/tigervnc/1.8.0/lib:/opt/munge-0.5.11/lib:/opt/slurm-17.11.4/lib:/opt/slurm-17.11.4/lib/slurm: Registering imageDump.mgz to hippocampal mask from ASEG $Id: mri_robust_register.cpp,v 1.77 2016/01/20 23:36:17 greve Exp $
--mov: Using imageDump.mgz as movable/source volume. --dst: Using /home/valentil/kg98/Valentina/ENIGMA_sex_differences/Scripts/rerun_reconall/NIAAA_MRI00032_1-out/tmp/hippoSF_T1_v21_left//hippoAmygBinaryMask_autoCropped.mgz as target volume. --lta: Output transform as trash.lta . --mapmovhdr: Will save header adjusted movable as imageDump_coregistered.mgz ! --sat: Using saturation 50 in M-estimator!
reading source 'imageDump.mgz'... reading target '/home/valentil/kg98/Valentina/ENIGMA_sex_differences/Scripts/rerun_reconall/NIAAA_MRI00032_1-out/tmp/hippoSF_T1_v21_left//hippoAmygBinaryMask_autoCropped.mgz'...
Registration::setSourceAndTarget(MRI s, MRI t, keeptype = TRUE ) Type Source : 0 Type Target : 3 ensure both FLOAT (3) Reordering axes in mov to better fit dst... ( -1 3 -2 ) MRIreorder() ----------- xdim=-1 ydim=3 zdim=-2 src 131 241 99, 0.250000 0.250000 0.250000 dst 131 99 241, 0.250000 0.250000 0.250000 Determinant after swap : 0.015625 Mov: (0.25, 0.25, 0.25)mm and dim (131, 99, 241) Dst: (1, 1, 1)mm and dim (36, 31, 52) Asserting both images: 1mm isotropic - reslicing Mov ... -- changing data type from 0 to 3 (noscale = 0)... -- Original : (0.25, 0.25, 0.25)mm and (131, 99, 241) voxels. -- Resampled: (1, 1, 1)mm and (36, 31, 61) voxels. -- Reslicing using cubic bspline MRItoBSpline degree 3 - reslicing Dst ... -- Original : (1, 1, 1)mm and (36, 31, 52) voxels. -- Resampled: (1, 1, 1)mm and (36, 31, 61) voxels. -- Reslicing using cubic bspline MRItoBSpline degree 3
Registration::computeMultiresRegistration - computing centroids - computing initial transform -- using translation info - Get Gaussian Pyramid Limits ( min size: 16 max size: -1 ) - Build Gaussian Pyramid ( Limits min steps: 0 max steps: 0 ) - Build Gaussian Pyramid ( Limits min steps: 0 max steps: 0 ) - initial transform: Ti = [ ... 1.0000000000000 0 0 0.8149113483238 0 1.0000000000000 0 0.6091444604181 0 0 1.0000000000000 -10.4099368843102 0 0 0 1.0000000000000 ]
- initial iscale: Ii =1
Resolution: 0 S( 36 31 61 ) T( 36 31 61 ) Iteration(f): 1 -- diff. to prev. transform: 8.43588 Iteration(f): 2 -- diff. to prev. transform: 6.13157 Iteration(f): 3 -- diff. to prev. transform: 5.20397 Iteration(f): 4 -- diff. to prev. transform: 5.47621 Iteration(f): 5 -- diff. to prev. transform: 8.03452 max it: 5 reached!
- final transform: Tf = [ ... 0.9476591624290 -0.1857780011274 0.2596702642998 -6.9596494990127 0.2461981066435 0.9430226622820 -0.2238185664945 2.3191313534194 -0.2032943780569 0.2760340426825 0.9394022584233 -11.1737673430040 0 0 0 1.0000000000000 ]
- final iscale: If = 1
********************************************************** * * WARNING: Registration did not converge in 5 steps! * Problem might be ill posed. * Please inspect output manually! * **********************************************************
Final Transform: Adjusting final transform due to initial resampling (voxel or size changes) ... M = [ ... -0.2369147906072 0.0649175660749 0.0464445002818 20.3444784868448 -0.0615495266609 -0.0559546416236 -0.2357556655705 36.6877108257354 0.0508235945142 0.2348505646058 -0.0690085106706 -14.6334717181848 0 0 0 1.0000000000000 ]
Determinant : -0.015625
writing output transformation to trash.lta ... converting VOX to RAS and saving RAS2RAS...
mapmovhdr: Changing vox2ras MOV header (to map to DST) ...
To check aligned result, run: freeview -v /home/valentil/kg98/Valentina/ENIGMA_sex_differences/Scripts/rerun_reconall/NIAAA_MRI00032_1-out/tmp/hippoSF_T1_v21_left//hippoAmygBinaryMask_autoCropped.mgz imageDump_coregistered.mgz
Registration took 0 minutes and 1 seconds.
Thank you for using RobustRegister! If you find it useful and use it for a publication, please cite:
Highly Accurate Inverse Consistent Registration: A Robust Approach M. Reuter, H.D. Rosas, B. Fischl. NeuroImage 53(4):1181-1196, 2010. http://dx.doi.org/10.1016/j.neuroimage.2010.07.020 http://reuter.mit.edu/papers/reuter-robreg10.pdf Highly Accurate Inverse Consistent Registration: A Robust ...http://reuter.mit.edu/papers/reuter-robreg10.pdf reuter.mit.edu Highly Accurate Inverse Consistent Registration: A Robust Approach Martin Reutera,b,c,∗, H. Diana Rosasa,b, Bruce Fischla,b,c aMassachusetts General Hospital / Harvard Medical School, Boston, MA, USA bMartinos Center for Biomedical Imaging, 143 13th Street, Charlestown, MA, USA cMIT Computer Science and AI Lab, Cambridge, MA, USA Abstract ...
$Id: mri_robust_register.cpp,v 1.77 2016/01/20 23:36:17 greve Exp $
--mov: Using imageDump.mgz as movable/source volume. --dst: Using /home/valentil/kg98/Valentina/ENIGMA_sex_differences/Scripts/rerun_reconall/NIAAA_MRI00032_1-out/tmp/hippoSF_T1_v21_left//hippoAmygBinaryMask_autoCropped.mgz as target volume. --lta: Output transform as trash.lta . --mapmovhdr: Will save header adjusted movable as imageDump_coregistered.mgz ! --affine: Enableing affine transform! --sat: Using saturation 50 in M-estimator!
reading source 'imageDump.mgz'... reading target '/home/valentil/kg98/Valentina/ENIGMA_sex_differences/Scripts/rerun_reconall/NIAAA_MRI00032_1-out/tmp/hippoSF_T1_v21_left//hippoAmygBinaryMask_autoCropped.mgz'...
Registration::setSourceAndTarget(MRI s, MRI t, keeptype = TRUE ) Type Source : 0 Type Target : 3 ensure both FLOAT (3) Reordering axes in mov to better fit dst... ( -1 3 -2 ) MRIreorder() ----------- xdim=-1 ydim=3 zdim=-2 src 131 241 99, 0.250000 0.250000 0.250000 dst 131 99 241, 0.250000 0.250000 0.250000 Determinant after swap : 0.015625 Mov: (0.25, 0.25, 0.25)mm and dim (131, 99, 241) Dst: (1, 1, 1)mm and dim (36, 31, 52) Asserting both images: 1mm isotropic - reslicing Mov ... -- changing data type from 0 to 3 (noscale = 0)... -- Original : (0.25, 0.25, 0.25)mm and (131, 99, 241) voxels. -- Resampled: (1, 1, 1)mm and (36, 31, 61) voxels. -- Reslicing using cubic bspline MRItoBSpline degree 3 - reslicing Dst ... -- Original : (1, 1, 1)mm and (36, 31, 52) voxels. -- Resampled: (1, 1, 1)mm and (36, 31, 61) voxels. -- Reslicing using cubic bspline MRItoBSpline degree 3
Registration::computeMultiresRegistration - computing centroids - computing initial transform -- using translation info - Get Gaussian Pyramid Limits ( min size: 16 max size: -1 ) - Build Gaussian Pyramid ( Limits min steps: 0 max steps: 0 ) - Build Gaussian Pyramid ( Limits min steps: 0 max steps: 0 ) - initial transform: Ti = [ ... 1.0000000000000 0 0 0.8149148880414 0 1.0000000000000 0 0.6090566088288 0 0 1.0000000000000 -10.4098892270167 0 0 0 1.0000000000000 ]
- initial iscale: Ii =1
Resolution: 0 S( 36 31 61 ) T( 36 31 61 ) Iteration(f): 1 -- diff. to prev. transform: 41.9486 Iteration(f): 2 -- diff. to prev. transform: 11.6817 Iteration(f): 3 -- diff. to prev. transform: 6.67838 Iteration(f): 4 -- diff. to prev. transform: 7.29474 Iteration(f): 5 -- diff. to prev. transform: 6.81314 max it: 5 reached!
- final transform: Tf = [ ... 0.7187784317010 -0.7039049689666 0.2089453505114 9.6512053486057 0.0746036190898 0.1293914974162 0.0117417560783 14.1524707759350 -0.2465560963278 -0.6917283813412 1.0760848577522 2.1884563417970 0 0 0 1.0000000000000 ]
- final iscale: If = 1
********************************************************** * * WARNING: Registration did not converge in 5 steps! * Problem might be ill posed. * Please inspect output manually! * **********************************************************
Final Transform: Adjusting final transform due to initial resampling (voxel or size changes) ... M = [ ... -0.1796946067577 0.0522363297057 0.1759762437984 14.8124989197474 -0.0186509047068 0.0029354397057 -0.0323478742666 20.2771618184274 0.0616390300951 0.2690212102028 0.1729321033528 -29.4307376768930 0 0 0 1.0000000000000 ]
Determinant : -0.00250544
Decompose into Rot * Shear * Scale :
Rot = [ ... -0.8007108168802 0.0158873366257 0.5988403629232 -0.5106339926748 0.5045973184270 -0.6961569304125 0.3132333207947 0.8632086299216 0.3959239168991 ]
Shear = [ ... 1.0000000000000 0.1745654451980 -0.3575927026144 0.2370458412539 1.0000000000000 0.6912992608189 -0.4065670242970 0.5788093472423 1.0000000000000 ]
Scale = diag([ 0.1727145993906 0.2345325414370 0.1963688302739 ])
writing output transformation to trash.lta ... converting VOX to RAS and saving RAS2RAS...
mapmovhdr: Changing vox2ras MOV header (to map to DST) ...
To check aligned result, run: freeview -v /home/valentil/kg98/Valentina/ENIGMA_sex_differences/Scripts/rerun_reconall/NIAAA_MRI00032_1-out/tmp/hippoSF_T1_v21_left//hippoAmygBinaryMask_autoCropped.mgz imageDump_coregistered.mgz
Registration took 0 minutes and 1 seconds.
Thank you for using RobustRegister! If you find it useful and use it for a publication, please cite:
Highly Accurate Inverse Consistent Registration: A Robust Approach M. Reuter, H.D. Rosas, B. Fischl. NeuroImage 53(4):1181-1196, 2010. http://dx.doi.org/10.1016/j.neuroimage.2010.07.020 http://reuter.mit.edu/papers/reuter-robreg10.pdf Highly Accurate Inverse Consistent Registration: A Robust ...http://reuter.mit.edu/papers/reuter-robreg10.pdf reuter.mit.edu Highly Accurate Inverse Consistent Registration: A Robust Approach Martin Reutera,b,c,∗, H. Diana Rosasa,b, Bruce Fischla,b,c aMassachusetts General Hospital / Harvard Medical School, Boston, MA, USA bMartinos Center for Biomedical Imaging, 143 13th Street, Charlestown, MA, USA cMIT Computer Science and AI Lab, Cambridge, MA, USA Abstract ...
Reading contexts of file /usr/local/freesurfer/devel-20180612/average/HippoSF/atlas/compressionLookupTable.txt Constructing image-to-world transform from header information (asegModCHA.mgz) Constructing image-to-world transform from header information (/home/valentil/kg98/Valentina/ENIGMA_sex_differences/Scripts/rerun_reconall/NIAAA_MRI00032_1-out/tmp/hippoSF_T1_v21_left/imageDump.mgz) Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Transforming points Reading contexts of file /usr/local/freesurfer/devel-20180612/average/HippoSF/atlas/compressionLookupTable.txt -------------- Making Left-Cerebral-Cortex map to reduced label 1 Making alveus map to reduced label 1 Making subiculum-body map to reduced label 1 Making subiculum-head map to reduced label 1 Making Hippocampal_tail map to reduced label 1 Making molecular_layer_HP-body map to reduced label 1 Making molecular_layer_HP-head map to reduced label 1 Making GC-ML-DG-body map to reduced label 1 Making GC-ML-DG-head map to reduced label 1 Making CA4-body map to reduced label 1 Making CA4-head map to reduced label 1 Making CA1-body map to reduced label 1 Making CA1-head map to reduced label 1 Making CA3-body map to reduced label 1 Making CA3-head map to reduced label 1 Making HATA map to reduced label 1 Making fimbria map to reduced label 1 Making presubiculum-body map to reduced label 1 Making presubiculum-head map to reduced label 1 Making parasubiculum map to reduced label 1 Making Lateral-nucleus map to reduced label 1 Making Paralaminar-nucleus map to reduced label 1 Making Basal-nucleus map to reduced label 1 Making Accessory-Basal-nucleus map to reduced label 1 Making Corticoamygdaloid-transitio map to reduced label 1 Making Central-nucleus map to reduced label 1 Making Cortical-nucleus map to reduced label 1 Making Medial-nucleus map to reduced label 1 Making Anterior-amygdaloid-area-AAA map to reduced label 1 -------------- Making Left-Cerebral-White-Matter map to reduced label 2 -------------- Making Left-Lateral-Ventricle map to reduced label 3 -------------- Making Left-choroid-plexus map to reduced label 4 -------------- Making hippocampal-fissure map to reduced label 5 Making Unknown map to reduced label 5 -------------- Making Left-VentralDC map to reduced label 6 -------------- Making Left-Putamen map to reduced label 7 -------------- Making Left-Pallidum map to reduced label 8 -------------- Making Left-Thalamus-Proper map to reduced label 9 -------------- Making Left-Accumbens-area map to reduced label 10 -------------- Making Left-Caudate map to reduced label 11 Smoothing mesh collection with kernel size 3.000000 ...Smoothing class: 0 Smoothing class: 1 Smoothing class: 2 Smoothing class: 3 Smoothing class: 4 Smoothing class: 5 Smoothing class: 6 Smoothing class: 7 Smoothing class: 8 Smoothing class: 9 Smoothing class: 10 done numberOfContrasts = 1 Image: 4 AtlasMeshToIntensityImage 1 0.398942 SLIDING L-BFGS Verbose: 0 MaximalDeformationStopCriterion: 1e-10 LineSearchMaximalDeformationIntervalStopCriterion: 1e-10 MaximumNumberOfIterations: 1000 BFGS-MaximumMemoryLength: 12 Resolution 1, iteration 1 Did one deformation step of max. 0 voxels in 4.2457 seconds
minLogLikelihoodTimesPrior =
2.4171e-16
Smoothing mesh collection with kernel size 2.000000 ...Smoothing class: 0 Smoothing class: 1 Smoothing class: 2 Smoothing class: 3 Smoothing class: 4 Smoothing class: 5 Smoothing class: 6 Smoothing class: 7 Smoothing class: 8 Smoothing class: 9 Smoothing class: 10 done numberOfContrasts = 1 Image: 4 AtlasMeshToIntensityImage 1 0.398942 SLIDING L-BFGS Verbose: 0 MaximalDeformationStopCriterion: 1e-10 LineSearchMaximalDeformationIntervalStopCriterion: 1e-10 MaximumNumberOfIterations: 1000 BFGS-MaximumMemoryLength: 12 Resolution 2, iteration 1 Did one deformation step of max. 0 voxels in 4.2086 seconds
minLogLikelihoodTimesPrior =
2.4171e-16
Fitting mesh to synthetic image from ASEG took 21.6669 seconds Error using segmentSubjectT1_autoEstimateAlveusML (line 533) Fitting mesh to synthetic image resulted in no deformation
@#@FSTIME 2018:09:25:10:18:56 run_segmentSubjectT1_autoEstimateAlveusML.sh N 13 e 65.81 S 1.73 U 34.89 P 55% M 1012356 F 271 R 219755 W 0 c 3155 w 11451 I 146312 O 477024 L 11.09 10.51 11.31 @#@FSLOADPOST 2018:09:25:10:20:02 run_segmentSubjectT1_autoEstimateAlveusML.sh N 13 35.39 16.50 13.25 Linux m3a007 3.10.0-514.26.1.el7.x86_64 #1 SMP Thu Jun 29 16:05:25 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
T1 hippocampal subfields exited with ERRORS at Tue Sep 25 10:20:02 AEST 2018
--
Sally Grace, BSc (Hons), PhD Candidate. Centre for Mental Health, Faculty of Health, Arts and Design Swinburne University of Technology