I am having troubles with registering the avgB0 to T1 image. Because some of the images are quite far off from each other, I am using an antsRegistrationSyN.sh call that includes an initial moving transform in the attempt of bringing them closer to one another:
Whereas for some subjects this yields a good result, for others the registration goes horribly wrong.
I thus tried to debug the problem by breaking down the call to check results of rigid only, and rigid+affine steps (always after initial moving transform). Here results vary again from subject to subject, being in some cases actually good and in others not perfect but also not horrible (i.e. nice registration in the posterior part of the brain, whereas the anterior part is still off).
In the cases where the rigid+affine steps give me a good result, I thus conclude that the problem stems from the deformable stage. I tried to increase the CC radius from 4 to 5 but still unsuccessful.
Do you have any idea of how I should modify my call to solve these issues? Or should I just consider stopping after rigid+affine?
Please put together a complete example that doesn't work as expected. Specifically, place both fixed and moving images in a folder along with a script with the antsRegistration command so I can simply download and run to begin diagnosing the problem.
Thanks,
Nick
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Please put together a complete example that doesn't work as expected. Specifically, place both fixed and moving images in a folder along with a script with the antsRegistration command so I can simply download and run to begin diagnosing the problem.
The -i option, as you're trying to use it here, is already included in the script call, i.e., you don't need to add this. The -i option, as it is is enabled in the script, is for when you have other transforms that have been previously calculated that you want to include (typically a warp file and/or a .mat file). Most use cases will typically not exercise this functionality.
Dear ANTs experts,
I am having troubles with registering the avgB0 to T1 image. Because some of the images are quite far off from each other, I am using an antsRegistrationSyN.sh call that includes an initial moving transform in the attempt of bringing them closer to one another:
antsRegistrationSyN.sh -d 3 -f T1BrainExtractionBrain.nii.gz -m skullstripped_avgB0.nii.gz -o b0toT1 -i [T1BrainExtractionBrain.nii.gz,skullstripped_avgB0.nii.gz,1]
Whereas for some subjects this yields a good result, for others the registration goes horribly wrong.
I thus tried to debug the problem by breaking down the call to check results of rigid only, and rigid+affine steps (always after initial moving transform). Here results vary again from subject to subject, being in some cases actually good and in others not perfect but also not horrible (i.e. nice registration in the posterior part of the brain, whereas the anterior part is still off).
In the cases where the rigid+affine steps give me a good result, I thus conclude that the problem stems from the deformable stage. I tried to increase the CC radius from 4 to 5 but still unsuccessful.
Do you have any idea of how I should modify my call to solve these issues? Or should I just consider stopping after rigid+affine?
Here's are the data of a subject for which the rigid+affine step goes well, and the SyN stage really doesn't.
https://drive.google.com/file/d/13DsdpQtc3HUjPlId3L82GYwV-ZW3XijB/view?usp=sharing
https://drive.google.com/file/d/1XTN0-ujLYEFBgGiAQNwDUW_0ciKXwQoe/view?usp=sharing
The T1 image has been processed with antsBrainExtraction.sh, whereas the DWI image is the skull-stripped average b0 after eddy currents and motion correction (done with FSL's eddy).
Any help would be greatly appreciated!
Thanks in advance,
Best,
Sam
Hi Sam,
Please put together a complete example that doesn't work as expected. Specifically, place both fixed and moving images in a folder along with a script with the
antsRegistrationcommand so I can simply download and run to begin diagnosing the problem.Thanks,
Nick
Hi Nick,
Thanks for your reply.
Here’s the folder I put together, let me know if you have problems with it or if you need any other file:
https://drive.google.com/open?id=1DTgDFPSsVawDPqXQisgWlgRAt79vScbs
Best,
Sam
From: Nick Tustison ntustison@users.sourceforge.net
Reply-To: "[advants:discussion]" 840260@discussion.advants.p.re.sourceforge.net
Date: Thursday, 23 May 2019 at 16:07
To: "[advants:discussion]" 840260@discussion.advants.p.re.sourceforge.net
Subject: [advants:discussion] Re: Registration issues with deformable stage in antsRegistrationSyN.sh
Hi Sam,
Please put together a complete example that doesn't work as expected. Specifically, place both fixed and moving images in a folder along with a script with the antsRegistration command so I can simply download and run to begin diagnosing the problem.
Thanks,
Nick
Registration issues with deformable stage in antsRegistrationSyN.shhttps://sourceforge.net/p/advants/discussion/840260/thread/fdf51cd20b/?limit=25#862a/559a
Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/advants/discussion/840260/
To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/
Hi Sam,
This syntax
The
-ioption, as you're trying to use it here, is already included in the script call, i.e., you don't need to add this. The-ioption, as it is is enabled in the script, is for when you have other transforms that have been previously calculated that you want to include (typically a warp file and/or a .mat file). Most use cases will typically not exercise this functionality.Try this call
It looks pretty good to me but let me know if you see something problematic.
Nick