[Elastix] Transform Rigidity Penalty

Jihun Kim jihun at umich.edu
Tue Apr 30 03:50:58 CEST 2013


I have looked into the related codes just a bit more in depth. It looks
that, specially when an initial transformation is defined, using Transform
Rigidity Penalty may do not work properly. This is because the origin of
m_RigidityCoefficientImage is shifted when an initial transformation is
applied while the origin of m_FixedRigidityImage is unchanged. Therefore,
when filling m_RigidityCofficientImage from m_FixedRigidityImage, there
would be some errors.

Regards,
Jihun
--
*Jihun Kim*
Ph. D. candidate
Mechanical Engineering
University of Michigan, Ann Arbor
734-276-7167


On Sun, Apr 28, 2013 at 5:26 AM, Coert Metz <coert.metz at gmail.com> wrote:

> Dear Jihun,
>
> The range of the MI metric is probably different from the range of the SSD
> metric, so I can imagine that you have to adapt a weighting parameter for
> the penalty term. Check the metric values in the output of both your MI and
> SSD registration to get an idea about an appropriate weighting factor.
>
> HTH,
> Coert Metz
>
>
> On Fri, Apr 26, 2013 at 11:55 PM, Jihun Kim <jihun at umich.edu> wrote:
>
>> Dear Elastix users,
>>
>> I have a question about the use of Transform Rigidity Penalty. I have
>> successfully used the penalty with Advanced Mean Squares, but when it is
>> coupled with Advanced Mutual Information I could not observe the effect of
>> the penalty terms. Could anybody give me some idea on this?
>>
>> Regards,
>> Jihun
>> --
>> *Jihun Kim*
>> Ph. D. candidate
>> Mechanical Engineering
>> University of Michigan, Ann Arbor
>> 734-276-7167
>>
>> _______________________________________________
>> Elastix mailing list
>> Elastix at bigr.nl
>> http://lists.bigr.nl/cgi-bin/mailman/listinfo/elastix
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bigr.nl/pipermail/elastix/attachments/20130429/7000a134/attachment.html>


More information about the Elastix mailing list