[Insight-users] Nonlinear registration of MRI to CT

[Marius Staring]

Hi Josh,

ITK is made for just that!

Because you want to register between two different modalities, you need a 
similarity measure that is robust for that. That is the most important 
thing. I have good experience with the mutual information similarity measure.

I usually use the itkMattesMutualInformationImageToImageMetric.

You can use any deformable transform for this task. For example a 
BSplineDeformableTransform can do the job.

I suggest you also use a multiresolution framework. You can study one of 
the examples in The Guide.

Regards,

Marius



At 09:27 20/07/04, Joshua Cates wrote:
>Hi all,
>
>Does anyone have experience with nonlinear registration of (T1) MRI onto
>CT?  Are any of the deformable algorithms in ITK up to the task?  Any
>suggestions would be appreciated.
>
>thanks,
>
>Josh
>
>
>
>---------- Forwarded message ----------
>Date: Mon, 19 Jul 2004 18:03:38 -0600
>From: Carsten Wolters <wolters at sci.utah.edu>
>To: insight-users at itk.org
>Cc: Seok Lew <slew at sci.utah.edu>, Joshua Cates <cates at sci.utah.edu>,
>      Darby J Van Uitert <darbyb at cs.utah.edu>
>Subject: Nonlinear registration
>
>Hi,
>
>I would like to register a T1-MRI onto a CT-image. Since the MR-image is
>nonlinear distorted, I would have to use a nonlinear registration approach.
>Are there methods in the ITK-toolbox that I could test for this purpose?
>Thanks in advance and cheers
>    carsten
>
>--
>Carsten H. Wolters, Ph.D.         Room 3493
>SCI Institute                     University of Utah
>50 S. Central Campus Dr.          Salt Lake City, UT 84112, USA
>Phone: (801) 587-7876             Fax:   (801) 585-6513
>http://www.sci.utah.edu/personnel/wolters.html
>
>
>
>
>_______________________________________________
>Insight-users mailing list
>Insight-users at itk.org
>http://www.itk.org/mailman/listinfo/insight-users