[Insight-users] Problem about the ITKAPP-IntensityBased2D3DRegistration : API Changes in CenteredEuler3DTra

[jimmy]

Dear Mr Luis Ibanez:

You really helped me a lot!  And I have learned much about Optimizer and Transform under your suggestion and guidance.
Thank you very much indeed!


jimmy
2007-07-24



 Luis Ibanez
 2007-07-23 21:20:38
 jimmy
 insight-users; Julien Jomier
 Re: Re: [Insight-users] Problem about the ITKAPP-IntensityBased2D3DRegistration : API Changes in CenteredEuler3DTra


Hi Jimmy,

Thanks for letting us know about your version of Insight and InsightApplications.

This seems to be a real bug:

The API of itkCenteredEuler3DTransforms changed around March 2005, and this
program was not updated correctly to match the new API.

http://www.itk.org/cgi-bin/viewcvs.cgi/Code/Common/itkCenteredEuler3DTransform.h?root=Insight&r1=1.4&r2=1.5 

The most recent API of the itkCenteredEuler3DTransform  used 9 parameters:

   3 for rotation
   3 for translation
   3 for the center of rotation

The program however, still expects only 6 parameters: 

  3 for rotation
  3 for translation


We have just reported this as Bug#5382:
http://public.kitware.com/Bug/bug.php?op=show&bugid=5382 


The simplest solution is for you to replace the

   CenteredEuler3DTransform

with the

   Euler3DTransform 

We are now making similar changes in the CVS version of InsightApplications. 


----


  Regarding the meaning of the Optimizer scales please read the 
  ITK Software Guide

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


   in particular the "Image Registration" chapter.

   The purpose of the scaling array that is passed to the optimizer 
   is to make more uniform the dynamic range of the Transform 
   parameters. E.g.  radians in rotations versus millimeters in 
   Translation.

   The size of the parameter scaling array *MUST* be the same 
   of the number of Transform parameters, since each element 
   of the parameter scaling array correspond to one of the 
   Transform parameters.
    

   This is explained in the ITK Software Guide, and has been 
   discussed many times in the list. Please read the guide and
   let us know if you still have any questions. 


       Thanks


            Luis




==============================================

On 7/23/07, jimmy < jimmy_lab at 163.com> wrote:
Thank you for your reply.

The version of Insight and InsightApplication are both 2.8.1.
The source file of IntensityBased2D3DRegistration is also attached.

Line 994 is where the problem located.
Could you tell me that what the meaning of weightings here used by Optimizer?
optimizer->SetScales(weightings);  // <- what do the scales stand for? 

I have also downloaded the newest version of IntensityBased2D3DRegistration, it seems to be unchanged.
The file version is  "Version:   $Revision: 1.4 $ ".

Thanks in advance.


------------------ 
jimmy
2007-07-23

-------------------------------------------------------------
发件人：Luis Ibanez
发送日期：2007-07-22 23:01:25
收件人：jimmy
抄送：insight-users
主题：Re: [Insight-users] Problem about the        ITKAPP-IntensityBased2D3DRegistration 


Hi Jimmy,

   What version of Insight             are you using ?
   What version of InsightApplications are you using ?

Over time the API of some of the ITK transforms has changed.
Notably, the Center of rotation, used to be part of the 
parameters of the transform. This may explain the mismatch
between 6 and 9 parameters.

It is likely that you are combining two imcompatible versions
of Insight and InsightApplications.


Please let us know, 


   Thanks


      Luis


--------------
jimmy wrote:
> Hi, everyone:
>
> When I running the ITKAPP [IntensityBased2D3DRegistration],
> an error occurs as the following: 
>
> itk::ExceptionObject (00F7FCD0)
> Location: "void __thiscall itk::RegularStepGradientDescentBaseOptimizer::Advance
> OneStep(void)"
> File: ITKSource\Code\Numerics\itkRegularStepGradientDescentBaseOptimizer.cxx 
> Line: 200
> Description: itk::ERROR: RegularStepGradientDescentOptimizer(015128A8): The size
>  of Scales is 6, but the NumberOfParameters for the CostFunction is 9.
>
> Can someone tell my why?   ( The test example is from BrainWeb ) 
>
> 1.
> Is it the problem of
> itk::Optimizer::ScalesType weightings( 6 );  ?
>
> What does it use for?  How to expand it into *9?*
>
> 2.
> And when I ignore the option of the command line 
>  ./IntensityBased2D3DRegistration [-t -100 -100 -100] projection.mhd /data/BrainWeb/brainweb165a10f17.mha
>
> The iteration of optimizer is 0, and the parameters of transform do not change.
> furthermore the programe crashed when writing the wrong result image. 
>
> Don't you think it is more reasonable that we should not input any tranformation information before registration?
>
>
>
> --------------
> jimmy
> 2007-07-18
> 
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/insight-users/attachments/20070724/bf131283/attachment-0001.html