[Insight-users] About the metric derivative of image registration

[Dan Mueller]

Hi Huanxiang,

> I’d like to know if there is an optimizer in ITK which is derivative free.
OnePlusOneEvolutionaryOptimizer does not use the gradient.


> Or is there any other way to calculate the derivative value?
Have you tried a simple finite difference approach?
For an example have a look at the function fdgrad(..)
      Utilities\vxl\core\vnl\vnl_cost_function.cxx

Hope this helps.

Regards, Dan

2009/5/5 Huanxiang Lu <huanxiang.lu at artorg.unibe.ch>:
> Hello,
>
>
>
> I am creating a new metric using CUDA implementation. So far I found it is
> very difficult to get the derivative value from my implementation. I’d like
> to know if there is an optimizer in ITK which is derivative free. Or is
> there any other way to calculate the derivative value?
>
>
>
> Thanks a lot!
>
>
>
> Huanxiang Lu
>
>
>
>
>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> 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
>
>