[prev in list] [next in list] [prev in thread] [next in thread] 

List:       insight-users
Subject:    [Insight-users] Multi-Resolution Demons with initial Affine
From:       Emma Saunders <emmasaunders123 () gmail ! com>
Date:       2013-07-31 8:26:16
Message-ID: CABQY8kTROzFR2t5K2RNptFfWCV03Bse8o1Hr1uR9j6NUFqTSEA () mail ! gmail ! com
[Download RAW message or body]

[Attachment #2 (multipart/alternative)]


Hi everyone,

I'm attempting some multi-resolution demons, as input I use an initial
affine transform.  When monitoring convergence however, I get peaks at
every pyramid level (see attached diagram).  Is this normal behaviour

 my pseudocode looks something like:

//Define Demons filter

 RegistrationFilterType::Pointer filter = RegistrationFilterType::New();

//Create initial Affine transform and convert to Displacement field, called
AffineField

  AffineField=converter->GetOutput();

  filter->SetInput(AffineField);
  filter->SetStandardDeviations( 0.8);

//use multiResolution

 MultiResRegistrationFilterType::Pointer multires =
MultiResRegistrationFilterType::New();

  multires->SetRegistrationFilter( filter );
  multires->SetNumberOfLevels( 4 );
  multires->SetFixedImage( FixedImage );
  multires->SetMovingImage( MovedImage );
  unsigned int nIterations[4] = {100, 100, 50, 50 };
  multires->SetNumberOfIterations( nIterations );
  multires->Update();

Am I correct in setting the initial affine field in this manner, does this
set the affine field for the first stage of the pyramid? or is it a case of
the affine field being re-applied at every pyramid value and hence peaks in
the metric

Thoughts appreciated

Thanks

Emma

[Attachment #5 (text/html)]

<div dir="ltr">Hi everyone,<div><br></div><div style>I&#39;m attempting some \
multi-resolution demons, as input I use an initial affine transform.  When monitoring \
convergence however, I get peaks at every pyramid level (see attached diagram).  Is \
this normal behaviour</div> <div style><br></div><div style> my pseudocode looks \
something like:<br></div><div style><br></div><div style>//Define Demons \
filter</div><div style><br></div><div style> RegistrationFilterType::Pointer filter = \
RegistrationFilterType::New();<br> </div><div style><br></div><div style>//Create \
initial Affine transform and convert to Displacement field, called \
AffineField</div><div style><br></div><div style><div>  \
AffineField=converter-&gt;GetOutput();</div><div>  </div> </div><div style><div><div> \
filter-&gt;SetInput(AffineField);</div><div>  filter-&gt;SetStandardDeviations( \
0.8);</div></div><div><br></div><div style>//use multiResolution</div><div \
style><br></div><div style><div> MultiResRegistrationFilterType::Pointer multires = \
MultiResRegistrationFilterType::New();</div> <div><br></div><div>  \
multires-&gt;SetRegistrationFilter( filter );</div><div>  \
multires-&gt;SetNumberOfLevels( 4 );</div><div>  multires-&gt;SetFixedImage( \
FixedImage );</div><div>  multires-&gt;SetMovingImage( MovedImage );</div> <div>  \
unsigned int nIterations[4] = {100, 100, 50, 50 };</div><div>  \
multires-&gt;SetNumberOfIterations( nIterations );</div><div style>  \
multires-&gt;Update();</div><div><br></div></div><div style>Am I correct in setting \
the initial affine field in this manner, does this set the affine field for the first \
stage of the pyramid? or is it a case of the affine field being re-applied at every \
pyramid value and hence peaks in the metric</div> <div style><br></div><div \
style>Thoughts appreciated</div><div style><br></div><div style>Thanks</div><div \
style><br></div><div \
style>Emma</div><div><br></div><div><br></div><div><br></div></div></div>

--047d7b160429ad4a3104e2ca7aaa--


["Iter.jpg" (image/jpeg)]

_____________________________________
Powered by www.kitware.com

Visit other Kitware open-source projects at
http://www.kitware.com/opensource/opensource.html

Kitware offers ITK Training Courses, for more information visit:
http://www.kitware.com/products/protraining.php

Please keep messages on-topic and check the ITK FAQ at:
http://www.itk.org/Wiki/ITK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.itk.org/mailman/listinfo/insight-users


[prev in list] [next in list] [prev in thread] [next in thread]