Hello (again!),
I would like to do a registration on two images : the reference is the RGB eigenvector and the source the T1 RMI.
The problem is that I don't find what I have to write in source_to_reference and reference_to_source...
What do I have to choose?
Thank you very much,
Aurélie.
Rigid Registration
-
riviere
- Site Admin
- Posts: 1361
- Joined: Tue Jan 06, 2004 12:21 pm
- Location: CEA NeuroSpin, Saint Aubin, France
- Contact:
Re: Rigid Registration
Hi,
Are you talking about the mutual information rigid registration process ?
source_to_reference and reference_to_source parameters are the output filenames for the transformation matrixes which will be estimated (in both directions). The process is too general to make use of the database for these parameters, so you have to specify/type filenames by hand (so, don't use the red database icons for them, but rather the file browser).
I'm not a specialist of registration (you can even say I don't know anything at all about registration), mutual information is a bit magic, but I'm not sure it will work well on these kind of images... Moreover I would bet it doesn't accept RGB images at all... If you want to register anatomical T1 and diffusion images, you'd better register the T1 and the T2-diffusion images, and the T2 should be already aligned to the diffusion because they should be acquired together.
Denis
Are you talking about the mutual information rigid registration process ?
source_to_reference and reference_to_source parameters are the output filenames for the transformation matrixes which will be estimated (in both directions). The process is too general to make use of the database for these parameters, so you have to specify/type filenames by hand (so, don't use the red database icons for them, but rather the file browser).
I'm not a specialist of registration (you can even say I don't know anything at all about registration), mutual information is a bit magic, but I'm not sure it will work well on these kind of images... Moreover I would bet it doesn't accept RGB images at all... If you want to register anatomical T1 and diffusion images, you'd better register the T1 and the T2-diffusion images, and the T2 should be already aligned to the diffusion because they should be acquired together.
Denis
-
aurely1310
- Posts: 38
- Joined: Tue Apr 06, 2010 10:22 am
Re: Rigid Registration
Hello Denis,
Yes, I'm talking about the mutual information rigid registration process.
It was what I did at the beginning, but the problem is that I would like (for the fiber tracking) define some ROI.
And I don't know on each image I have to do that : ideally i would like to define the ROI on the T1 image.
But when the fibers are constructed, which model would be taken in account ?? FOr example :
(I used the "draw ROI" button from brainvisa)
- I drew ROI on T1 (filled in each field : the ROI are saved)
- Next, instead to open the T1 image, I opened the T2 image. And here the "old-defined" regions are completely false!
That's why I'm a little bit lost...
Could you help me please? (And thank you for all your replies...!)
Yes, I'm talking about the mutual information rigid registration process.
It was what I did at the beginning, but the problem is that I would like (for the fiber tracking) define some ROI.
And I don't know on each image I have to do that : ideally i would like to define the ROI on the T1 image.
But when the fibers are constructed, which model would be taken in account ?? FOr example :
(I used the "draw ROI" button from brainvisa)
- I drew ROI on T1 (filled in each field : the ROI are saved)
- Next, instead to open the T1 image, I opened the T2 image. And here the "old-defined" regions are completely false!
That's why I'm a little bit lost...
Could you help me please? (And thank you for all your replies...!)
-
Dominique Geffroy
- Site Admin
- Posts: 161
- Joined: Thu Mar 01, 2007 4:22 pm
- Location: IFR 49 - Neurospin, Gif-sur-Yvette, France
- Contact:
Re: Rigid Registration
Hello Aurely,
You can use ROI drawn on the T1 image as the starting points of the tracking. But in that case, you have to fill in the parameter starting_points_transformation parameter in fiber tracking process with the transformation between the T1 image and the diffusion image. To compute that transformation, maybe it is more practical to use the process AC-PC Alignment -> Diffusion to AC/PC referential in Diffusion and Tracking toolbox, it uses also mutual information to compute the transformation between T1 and T2 images, but it uses the database features, so the filenames are automatically filled in.
Dominique
You can use ROI drawn on the T1 image as the starting points of the tracking. But in that case, you have to fill in the parameter starting_points_transformation parameter in fiber tracking process with the transformation between the T1 image and the diffusion image. To compute that transformation, maybe it is more practical to use the process AC-PC Alignment -> Diffusion to AC/PC referential in Diffusion and Tracking toolbox, it uses also mutual information to compute the transformation between T1 and T2 images, but it uses the database features, so the filenames are automatically filled in.
Dominique