Hi Andrew,
Sorry for the late response. Copied below is the suggestions from Martin.
“ I looked at those images. A few points:
- bias fields seem to be very different, I think using --iscale helps a little here, if very high accuracy of the registration is important, one should do a bias field removal beforehand.
- manually setting the sat to anything between 5 and 8 also looked like it helps --sat 5 in robust register
- robust register can be used if only two time points are involved to register into a mid space using —halfdst halfmov and halfdstlta outputs you get those images and transforms. Only thing missing is the actual template image which for two time points is simply the average of the two images after alignment in the mid space (so probably mri_calc or some other tool can compute that average image if it is needed).
- Scalp is changing a lot here across time, not sure why. How far are these time points apart? What type of data (aging, disease, treatment) is this? There are also noticeable changes in the brain. Anyway, best results are when using the skull stripped images, it is also faster. In freesurfer we work with norm.mgz in the longitudinal pipeline which are both skull stripped an bias corrected.
- Finally, in robust_template I also get the error about deviation from rotation. For some reason, that I could not yet find out, the first registration of A to B is not perfectly a rigid transform (some small deviation from a rotation matrix, maybe very subtle shearing as the determinant is still very close to 1, so no scaling). This could be a numerical issue. Anyway increasing the threshold to 0.001 should fix this. Or did you get another error after that for this specific case? “
From: Andrew Beaudin abeaudin@ucalgary.ca Sent: Wednesday, December 7, 2022 1:30 PM To: Huang, Yujing YHUANG43@mgh.harvard.edu; Freesurfer support list freesurfer@nmr.mgh.harvard.edu Cc: Reuter, Martin,Ph.D. MREUTER@mgh.harvard.edu Subject: Re: [Freesurfer] Fwd: mri_robust_template - MyMatrix::RotationMean Error
External Email - Use Caution Hi Yujing, No problem at all. I really appreciate yours, and now Martin's, help with this issue. If you would like to start with dicoms, please let me know.
Again, thank you both for your help.
Andrew
---- Andrew Beaudin, PhD | Research Associate University of Calgary | Cumming School of Medicine Department of Clinical Neurosciences HSC Rm 2935J, 3330 Hospital Drive N.W. | Calgary, AB | T2N 4N1 Phone: (403) 210-6273
________________________________ From: Huang, Yujing [mailto:YHUANG43@mgh.harvard.edu] Sent: Wednesday, December 7, 2022 at 10:27 To: Andrew Beaudin abeaudin@ucalgary.camailto:abeaudin@ucalgary.ca, Freesurfer support list freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu Cc: Reuter, Martin,Ph.D. MREUTER@mgh.harvard.edumailto:MREUTER@mgh.harvard.edu Subject: [Freesurfer] Fwd: mri_robust_template - MyMatrix::RotationMean Error
[△EXTERNAL]
Hi Andrew,
Sorry for the late response. I’m able to reproduce your error.
I have included Martin to the communication. He is the author of the program. He is going to take a look at your data.
Best,
Yujing
From: Andrew Beaudin abeaudin@ucalgary.camailto:abeaudin@ucalgary.ca Sent: Friday, November 25, 2022 3:25 PM To: Huang, Yujing YHUANG43@mgh.harvard.edumailto:YHUANG43@mgh.harvard.edu; Freesurfer support list freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu Subject: RE: [Freesurfer] Fwd: mri_robust_template - MyMatrix::RotationMean Error
External Email - Use Caution Hi Huang, Thank you very much for your help with this. Apologies for the delay.
You should be able to retrieve sample images from the following link (expires on Dec. 2). I have also included a brief word document at the link that includes some details of the nifti images.
https://uofc-my.sharepoint.com/:f:/g/personal/abeaudin_ucalgary_ca/EmL7JGE0J...https://secure-web.cisco.com/1hstJ7MT7lbWknwArluI7lsJDpCaPEnPjFOntWZC705TN7hf2zZ6yi6kYZOC0uvnzY4lhzK3ZM6JenT-iUCVRXmGQD3uftAD8OYL8nRsUntQSgd62pdkysnwz7Ei0MpfFScbOiKSm0zYOh7orQWmOfvRZJYI8GvWgu9XtDcLuLHMfgMkdllTtYJdE5Ht3Rh2a17I1tqEXZCjpKzzl_Hdx6NaDHJPqGxeKLX0pSSO_LGk3WyZ7S3YL04EgD6GiMRV2KCYSn4mx2cLl7I06xdwOr9DxyUFAXoLsVoBpaUDi-KcISE065Vag1H-7HyUhCIAgQFmsEVYjvotyAmUOYZvUQw/https%3A%2F%2Fuofc-my.sharepoint.com%2F%3Af%3A%2Fg%2Fpersonal%2Fabeaudin_ucalgary_ca%2FEmL7JGE0J-BHpnRfqArO3OMBlcD4VPqRkUZKfsmKGeL2SQ
The command call I am using is:
mri_robust_template --mov 2001A_T1o_brain.nii.gz 2001B_T1o_brain.nii.mgz --lta 2001A2template.lta 2001B2template.lta --average 1 --template 2001_template.nii.gz --satit
However, while troubleshooting I have also tried the following with no success:
1) using the --iscale flag - same MyMatrix::RotationMean Error (see below for log) 2) using --inittp 0 - results in a 'Segmentation fault' 3) using --average 0 - same MyMatrix::RotationMean Error 4) using Freesurfer v7.2 - same MyMatrix::RotationMean Error 5) running mri_robust_template call on Catalina and Big Sur systems with FS v6.0 - same MyMatrix::RotationMean Error.
Again, thank you for the help….and please let me know if your need any additional information or anything clarified.
Andrew ---- Andrew Beaudin, PhD | Postdoctoral Scholar University of Calgary | Cumming School of Medicine Department of Clinical Neurosciences HSC Rm 2935J, 3330 Hospital Drive N.W. | Calgary, AB | T2N 4N1 Phone: (403) 210-6273
From: Huang, Yujing <YHUANG43@mgh.harvard.edumailto:YHUANG43@mgh.harvard.edu> Sent: November 22, 2022 7:32 AM To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu> Cc: Andrew Beaudin <abeaudin@ucalgary.camailto:abeaudin@ucalgary.ca> Subject: RE: [Freesurfer] Fwd: mri_robust_template - MyMatrix::RotationMean Error
[△EXTERNAL]
Hi Andrew,
Can you send us some data to debug? The link has expired. Also, the command line that you are having trouble with.
Thanks.
Yujing
From: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu <freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu> On Behalf Of Andrew Beaudin Sent: Wednesday, June 29, 2022 10:45 PM To: freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Fwd: mri_robust_template - MyMatrix::RotationMean Error
External Email - Use Caution Hello Freesurfer Community, I am trying to use mri_robust_template (with its default 6 DOF registration) outside of the Freesurfer longitudinal pipeline to generate within-participant T1 templates from 2 time points. Unfortunately, I am coming up against a 'MyMatrix::RotationMean input(s) not rotation' error on images acquired on our GE 3T Discovery 750 system, but not our colleagues' Seimens 3T Prisma images. I have pasted the output from the --debug flag below (and attached as text file - line 188 is the MyMatrix message). If helpful, images can be download from the following link: MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://uofc-my.sharepoint.com/:f:/g/personal/abeaudin_ucalgary_ca/EmL7JGE0J...https://secure-web.cisco.com/1FGQFgBoMNs00DgWtLCrKSIIdug6juLkb2Nu2KrASFJPZBjdCx_joSq08q6GYn899WbiMHITtkL3x-o5AHvRXbH-2JHAu0_Gw5aFmg9rIhsPy0Zu8zpryZYHjsE4umwwEylBRyd0TGNV3m-d5k9_IsQhRu2SYc6vj4WIaJ8F_q9uIcsgpX_QVuppEMG4lRVPp3RvveMfZRpMkehNDZJ_pJT9pje_N5viL2Zb7XZPRfaNnCKK7hTcevtyaKDqtmy85Rm2GqvtUB9vFnPFssxAMeZH7UBH2-8RQCzhCokbVocDQ7-m0q6rQ-kCKxQSNMpVdE3gMPmdjbjJ6vIqZ0RuPLkj9Labe95_FuZXUxmFfL8Y/https%3A%2F%2Fuofc-my.sharepoint.com%2F%3Af%3A%2Fg%2Fpersonal%2Fabeaudin_ucalgary_ca%2FEmL7JGE0J-BHpnRfqArO3OMBlcD4VPqRkUZKfsmKGeL2SQ%3Fe%3DlcovfR
Briefly: - the images are skull-stripped niftis and I am running Freesurfer 6.0 (freesurfer-Linux-centos6_x86_64-stable-pub-v6.0.0-2beb96c) on a WSL2 Ubuntu 18.04 installation. - the call is: mri_robust_template --mov 2001A_T1o_brain.nii.gz 2001B_T1o_brain.nii.mgz --lta 2001A2template.lta 2001B2template.lta --average 1 --template 2001_template.nii.gz --satit
- I have tried the following with no success: 1) using the --iscale flag - same result as shown below. 2) using --inittp 0 - results in a 'Segmentation fault' 3) using --average 0 - same result as shown below 4) using Freesurfer v7.2 - same result as below 5) running mri_robust_template call on Catalina and Big Sur systems with FS v6.0 - same result as below.
Currently, the only way to get mri_robust_template to work with our GE images is to run it with the --affine flag. I would prefer to use the rigid body registration but am at a loss for why these images are failing using the rigid body registration is failing. Any help/insights/recommendations are greatly appreciated.
Thank you very much.
Andrew
---- Andrew Beaudin, PhD | Postdoctoral Scholar University of Calgary | Cumming School of Medicine Department of Clinical Neurosciences HSC Rm 2935J, 3330 Hospital Drive N.W. | Calgary, AB | T2N 4N1 Phone: (403) 210-6273
DEBUG REPORT:
$Id: mri_robust_template.cpp,v 1.54 2016/05/05 21:17:08 mreuter Exp $
--mov: Using 2001A_T1o_brain.nii.gz as movable/source volume. --mov: Using 2001B_T1o_brain.nii.gz as movable/source volume. Total: 2 input volumes --lta: Will output LTA transforms --average: Using method 1 for template computation. --template: Using 2001_template.nii.gz as template output volume. --satit: Will estimate SAT iteratively! --debug: Will output debug info and files! reading source '2001A_T1o_brain.nii.gz'... converting source '2001A_T1o_brain.nii.gz' to bspline ... MRItoBSpline degree 3 reading source '2001B_T1o_brain.nii.gz'... converting source '2001B_T1o_brain.nii.gz' to bspline ... MRItoBSpline degree 3 Will use TP 2 as random initial target (seed 108242 ).
MultiRegistration::initializing Xforms (init 2 , maxres 0 , iterate 5 , epsit 0.01 ) :
MultiRegistration Parameters
Outdir: ./ TransOnly: false Rigid: true Robust: true Satit: true Iscale: false IscaleOnly: false NoMulti: false SubsampleSize: -1 HighIt: -1 FixVoxel: false KeepType: true Average: 1 DoublePrec: false BackupWeights: false SampleType: 5 CRASCenter: false Debug: 1
[init] ========================= TP 1 to TP 2 ============================== Register TP 1 ( 2001A_T1o_brain.nii.gz ) to TP 2 ( 2001B_T1o_brain.nii.gz )
Registration::setSourceAndTarget(MRI s, MRI t, keeptype = TRUE ) Type Source : 3 Type Target : 3 keeping type Mov: (1, 1, 1)mm and dim (176, 256, 256) Dst: (1, 1, 1)mm and dim (176, 256, 256) Asserting both images: 1mm isotropic - no Mov reslice necessary - no Dst reslice necessary
Registration::findSaturation - 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: 3 ) - Build Gaussian Pyramid ( Limits min steps: 0 max steps: 3 )
- Max Resolution used: 2 -- gpS ( 44 , 64 , 64 ) -- gpT ( 44 , 64 , 64 ) - running loop to estimate saturation parameter: -- Iteration: 1 trying sat: 16 min sat: 16 ( 0.729941 ), max sat: 0 ( -1 ), sat diff: -16, (wlimit=0.16) -- Iteration: 2 trying sat: 32 min sat: 32 ( 0.713406 ), max sat: 0 ( -1 ), sat diff: -32, (wlimit=0.16) -- Iteration: 3 trying sat: 64 min sat: 64 ( 0.692689 ), max sat: 0 ( -1 ), sat diff: -64, (wlimit=0.16) -- Iteration: 4 trying sat: 128 min sat: 128 ( 0.666858 ), max sat: 0 ( -1 ), sat diff: -128, (wlimit=0.16) -- Iteration: 5 trying sat: 256 min sat: 256 ( 0.637119 ), max sat: 0 ( -1 ), sat diff: -256, (wlimit=0.16) -- Iteration: 6 trying sat: 512 min sat: 512 ( 0.591624 ), max sat: 0 ( -1 ), sat diff: -512, (wlimit=0.16) -- Iteration: 7 trying sat: 1024 min sat: 1024 ( 0.550606 ), max sat: 0 ( -1 ), sat diff: -1024, (wlimit=0.16) -- Iteration: 8 trying sat: 2048 min sat: 2048 ( 0.485668 ), max sat: 0 ( -1 ), sat diff: -2048, (wlimit=0.16) -- Iteration: 9 trying sat: 4096 min sat: 4096 ( 0.370434 ), max sat: 0 ( -1 ), sat diff: -4096, (wlimit=0.16) -- Iteration: 10 trying sat: 8192 min sat: 8192 ( 0.19981 ), max sat: 0 ( -1 ), sat diff: -8192, (wlimit=0.16) -- Iteration: 11 trying sat: 16384 min sat: 8192 ( 0.19981 ), max sat: 16384 ( 0.0606314 ), sat diff: 8192, (wlimit=0.16) -- Iteration: 12 trying sat: 12288 min sat: 8192 ( 0.19981 ), max sat: 12288 ( 0.103499 ), sat diff: 4096, (wlimit=0.16) -- Iteration: 13 trying sat: 10240 min sat: 8192 ( 0.19981 ), max sat: 10240 ( 0.142423 ), sat diff: 2048, (wlimit=0.16) -- Iteration: 14 trying sat: 9216 min sat: 9216 ( 0.16727 ), max sat: 10240 ( 0.142423 ), sat diff: 1024, (wlimit=0.16) -- Iteration: 15 trying sat: 9728 min sat: 9216 ( 0.16727 ), max sat: 9728 ( 0.155264 ), sat diff: 512, (wlimit=0.16) -- Iteration: 16 trying sat: 9472 min sat: 9472 ( 0.160488 ), max sat: 9728 ( 0.155264 ), sat diff: 256, (wlimit=0.16) -- Iteration: 17 trying sat: 9600 min sat: 9472 ( 0.160488 ), max sat: 9600 ( 0.157199 ), sat diff: 128, (wlimit=0.16) -- Iteration: 18 trying sat: 9536 min sat: 9472 ( 0.160488 ), max sat: 9536 ( 0.159131 ), sat diff: 64, (wlimit=0.16) -- Iteration: 19 trying sat: 9504 min sat: 9472 ( 0.160488 ), max sat: 9504 ( 0.159604 ), sat diff: 32, (wlimit=0.16) -- Iteration: 20 trying sat: 9488 min sat: 9472 ( 0.160488 ), max sat: 9488 ( 0.159903 ), sat diff: 16, (wlimit=0.16) -- Iteration: 21 trying sat: 9480 min sat: 9480 ( 0.160034 ), max sat: 9488 ( 0.159903 ), sat diff: 8, (wlimit=0.16) -- Iteration: 22 trying sat: 9484 min sat: 9484 ( 0.160123 ), max sat: 9488 ( 0.159903 ), sat diff: 4, (wlimit=0.16) -- Iteration: 23 trying sat: 9486 min sat: 9484 ( 0.160123 ), max sat: 9486 ( 0.159831 ), sat diff: 2, (wlimit=0.16) -- Iteration: 24 trying sat: 9485 min sat: 9484 ( 0.160123 ), max sat: 9485 ( 0.159908 ), sat diff: 1, (wlimit=0.16) -- Iteration: 25 trying sat: 9484.5 - final SAT: 9485 ( it: 25 , weight check 0.159908 <= 0.16 )
Registration::computeMultiresRegistration - computing centroids - computing initial transform -- using translation info - Get Gaussian Pyramid Limits ( min size: 16 max size: -1 ) - initial transform: Ti = [ ... 1.0000000000000 0 0 -2.0251428914711 0 1.0000000000000 0 6.0311902437196 0 0 1.0000000000000 2.8105113695639 0 0 0 1.0000000000000 ]
- initial iscale: Ii =1
Resolution: 3 S( 22 32 32 ) T( 22 32 32 ) Iteration(f): 1 -- diff. to prev. transform: 2.97694 Iteration(f): 2 -- diff. to prev. transform: 0.0903233 Iteration(f): 3 -- diff. to prev. transform: 0.0191757 Iteration(f): 4 -- diff. to prev. transform: 0.0116139 Iteration(f): 5 -- diff. to prev. transform: 0.00140989 <= 0.01 :-)
Resolution: 2 S( 44 64 64 ) T( 44 64 64 ) Iteration(f): 1 -- diff. to prev. transform: 1.81202 Iteration(f): 2 -- diff. to prev. transform: 0.199033 Iteration(f): 3 -- diff. to prev. transform: 0.00925469 <= 0.01 :-)
Resolution: 1 S( 88 128 128 ) T( 88 128 128 ) Iteration(f): 1 -- diff. to prev. transform: 1.02247 Iteration(f): 2 -- diff. to prev. transform: 0.0229168 Iteration(f): 3 -- diff. to prev. transform: 0.000719712 <= 0.01 :-)
Resolution: 0 S( 176 256 256 ) T( 176 256 256 ) Iteration(f): 1 -- diff. to prev. transform: 0.664624 Iteration(f): 2 -- diff. to prev. transform: 0.0108463 Iteration(f): 3 -- diff. to prev. transform: 0.000585613 <= 0.01 :-)
- final transform: Tf = [ ... 0.9994709868827 -0.0113151175724 -0.0304912199489 4.8530791096303 0.0126622658755 0.9989355014916 0.0443568583483 2.1370162587100 0.0299568590234 -0.0447194809223 0.9985503265352 5.4716006804225 0 0 0 1.0000000000000 ]
- final iscale: If = 1 CentroidT = [ 86.5860705940972 104.6067830299094 138.6676688710562 1.0000000000000 ]
CentroidSinT = [ 88.1595712077722 107.7558909090270 139.3780739549505 1.0000000000000 ]
Centroid = [ 87.3728209009347 106.1813369694682 139.0228714130033 1.0000000000000 ]
MyMatrix::RotationMean input(s) not rotation meant = [ -1.2248131651588 1.2892456671960 3.1420193014715 ]
Input rotation's max deviation from rotation is: 0.000393411
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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/compliancelinehttps://secure-web.cisco.com/1YFv6ynnQtVwxD61_92Xy0zSgOn5A4uR1nLgV2tl4XDmF3jch_6hkkwF6MtuN2LTDjOqfhiQ3Wg-yTmkMyfERS-4kdXSDkmxiWGEiNUogrRchZnezlAtCn_oR8CAnlXnLYv2Fp3B8GbRvrhIYEHc1Ti0WoQHCuhhyXFavIkMkewZYwDPvvmbDSrI5doE8xlT7O9uPQRQXX_o2uQuN9lgzLiu8lgCrEUm_pHncUNtZLqd965cSX31MiDapVNgx7zube-Ez5qRRLt_-EYuiRpfYxR-CpSVJVv2tooEf3K3H-1TQ2x_x8Td3rdvlMXC9vUdBXNqqIhU6yJTMVKJ5pmNsMw/https%3A%2F%2Fwww.massgeneralbrigham.org%2Fcomplianceline .
Please note that this e-mail is not secure (encrypted). If you do not wish to continue communication over unencrypted e-mail, please notify the sender of this message immediately. Continuing to send or respond to e-mail after receiving this message means you understand and accept this risk and wish to continue to communicate over unencrypted e-mail. 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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/compliancelinehttps://secure-web.cisco.com/1cUNAiKH4ulUP3fRKKP1t87y2Y7lfv51xZKoMD0jD3rUhfQTVmd-igLxTi0Cl3liUlJtaEb7pjHbfzFEg5wwui-kxZVgYSZJ_hQUokVv1PqGSaLXa9nIVIhBhBcVhAT_FGTZjL9sjWBYbLrHW0q4sfHZmPT_FEl5CUqPnBTCpY7je3JtiZD2Ch8K0H5KNo4BQBwdvAfw5qvpLdvNZ1GpgLXTcDBSK50G4-EE9ktbgUHm3zDDV972tnQMJvp49eF_1wL-B6hq4Ndtqh2HB9V9usE1FvvQD9n6qluTVCJtOXW5XdNKCtjzn5kmg9qdbxEtC9P_KaYUQrlJgj4Cl3gntIQ/https%3A%2F%2Fwww.massgeneralbrigham.org%2Fcomplianceline .
Please note that this e-mail is not secure (encrypted). If you do not wish to continue communication over unencrypted e-mail, please notify the sender of this message immediately. Continuing to send or respond to e-mail after receiving this message means you understand and accept this risk and wish to continue to communicate over unencrypted e-mail.