[Insight-users] registration pipeline suggestion

[Luis Ibanez]

Hi Rodrigo,


Is your functional dataset a fMRI image or a PET scan ?


In general terms you may want to use a combination such as:

A) METRIC
Mutual Information metrics in ITK:
http://www.itk.org/Insight/Doxygen/html/group__RegistrationMetrics.html

B) TRANSFORM AND OPTIMIZER
If you are registering a subject with itself you should
use the Rigid Transform:
http://www.itk.org/Insight/Doxygen/html/classitk_1_1VersorRigid3DTransform.html
and its associated optimizer
http://www.itk.org/Insight/Doxygen/html/classitk_1_1VersorRigid3DTransform.html

If you are registering two different subjects you
probably want the SimilarityTransform, or the AffineTransform.
http://www.itk.org/Insight/Doxygen/html/classitk_1_1Similarity3DTransform.html
http://www.itk.org/Insight/Doxygen/html/classitk_1_1AffineTransform.html
Note that the Affine transform is more expensive to compute
than the Similarity transform, and that functional datasets
may not have the anatomical information that will justify
deformations such as Shearing.


C) INTERPOLATOR
A linear interpolator is in general the best compromise
between precision and performance.
http://www.itk.org/Insight/Doxygen/html/classitk_1_1LinearInterpolateImageFunction.html
However if you are dealing with low resolution datasets you
may affort a more time-expensive interpolator such as BSplines.
http://www.itk.org/Insight/Doxygen/html/classitk_1_1BSplineInterpolateImageFunction.html


Also, you may want to give it a try to Slicer,

             http://www.slicer.org/intro/

Which already has a module for registration functional MRI based
on ITK methods. Slicer is free and Open Source.



    Regards,


       Luis



-------------------------------
rvivanco at cs.umanitoba.ca wrote:
> Hello,
> 
> The registration chapter of the user's guide is quite helpful and extensive, but
> with the many options for registration I was wondering if someone could give me
> a suggestion on which registration pipeline would work most of the times for
> registering a functional dataset, by itself and also accross subjects.
> 
> thanks,
> rodrigo
> 
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
> 
>