Thank you, we will try this.
In reality it seems not obvious if image A or B is the distorted one.
I wonder if we could use this for detecting absolute distortions in
parallel beam tomography. Two images acquired 180 degrees apart should
be mirror of each other. Discrepancies would be most likely related to
(optical) distortions. Maybe the requirement that the images, once
undistorted, should be mirror helps.
An image of a reference object like a grid also helps probably.
Peter
On Fri, 2006-12-01 at 12:07 +0100, Ignacio Arganda Carreras wrote:
> Dear all,
>
> I just uploaded a new stable version of the image registration plugin
> bUnwarpJ. It corrects some bugs from the previous release:
>
>
http://biocomp.cnb.csic.es/%7Eiarganda/bUnwarpJ/index.html>
> I called it a "beta" version because there is not too much
> documentation. I'll try to write a manual soon.
>
> There are also some new tools that allow comparing the algorithm to any
> other registration technique.
>
> Enjoy!
>
> ignacio
>