[Insight-users] Example ImageRegistration8

[Luis Ibanez]

Hi Ines,

A displacement of 26 pixels in both X and Y is a mayor shift,
this is unlikely to be corrected without using multi-resolution.   :-/



Here are two suggestions:



  A) Please build the example in InsightApplications:

          InsightApplications/Auxiliary/vtk/
                     itkRegisterITKMonitorVTK.cxx

     and run it on your two datasets.

     This example will show you the progress of the registration
     by using VTK surface renderings of your images. In this way
     you will know what is happening in 3D during the intermediate
     steps of the registration.


     BTW: What is the content of your images ?
          The surface extraction is simply using iso-surfaces,
          since they are only for the purpose of visualizing
          the progress of the registration.


AND/OR


B) Start your experiment with a displacement smaller than 26 pixels.

    E.g start with only 5 pixels only in X and see if the registration
    code can correct for it. If the registration succes, then start the
    experiment again with a larger misregistration... and so on. In this
    way you will identify the capture radius of the metric for your
    particular image.


    BTW: Please not that the registration process is done in physical
         units, not pixels.

         What is the pixel spacing of your images ?

         How are you passing that "26 pixels" displacement to the
         registration method ?



   Please let us know what you find.



      Thanks


         Luis



------------------------
ines cabezas rivas wrote:
> Hi all. I used the imageregistration8.cxx like the guide to work with 3D 
> model images.
> But I used the affine transform, the mean squares metric, the regular 
> step gradient optimizer and the lineal interpolator. I tested the 
> program and I used a serie like fixed image and the same serie out of 
> place 26 pixels in axis X and Y, so the results must be similar to the 
> displacement used, but the results are very diferents. The translation 
> obtained is similat to [0.1,0.27,-0.34] -GetLastTransformParameters()- 
> In the mean squares, i don´t configure the parameters. The parameters 
> that I configured are in the optimizer:
> maximumStepLenght = 0.5;
> minimumStepLenght = 1E-5;
> noIterations = 50.0;
> relaxionFactor = 0.8;
> 
> Does anyone have a idea, please?
> Thaks you..
> Best Regards!!
> 
> ------------------------------------------------------------------------
> Tecnología, moda, motor, viajes,.suscríbete a nuestros boletines para 
> estar siempre a la última MSN Newsletters 
> <http://newsletters.msn.com/hm/maintenanceeses.asp?L=ES&C=ES&P=WCMaintenance&Brand=WL&RU=http%3a%2f%2fmail.live.com>
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users