[Insight-users] Maximization of MI in ITK

[Luis Ibanez]

Hi Suresh,


Please read the "Image Registration" Chapter in the
ITK Software Guide


       http://www.itk.org/ItkSoftwareGuide.pdf



1) For Mattes MI you should concentrate on the parameters

     - Number of Histogram bins
     - Number of Spatial Samples

    You must profile the behavior of the Metric for known
    displacements. In this way, you will be able to appreciate
    the effect of the parameters on the final value of the
    Metric.


2) There are five different techniques for deformable
    registration in ITK

     - Demons
     - BSplines
     - FEM
     - Level Sets
     - Kernel Splines

    If the edges are being smeared, that means that you
    are not setting enough regularization. In the case
    of Demons, regularization is controled by the smoothing
    of the deformation field. Simply increase the smoothing,
    and that will make the deformation field more uniform.



3) Yes, FEM is implemented in ITK

    In FEM the regularization is controled with the
    Elasticity and Density parameters.




   Regards,


      Luis




-----------------
Suresh G N wrote:
> Dear All,
> 
> I am using Mattes MI metric with Translational transform to register my 
> dataset. Can some one tell me
> 
> 1) what are the parameters I have to consider carefully to optimize the 
> metric.
> 2) Is there any other deformable technique available in ITK, so that the 
> images edges can be retained correctly(Demon's gives smeared effect on 
> the image).
> 3) What about FEM based deformable registration?
> 
> thanks and regards
> Suresh
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users