Login  Register

Re: Autocorrelation using Fourier

Posted by Albert Cardona-2 on Dec 15, 2010; 3:19am
URL: http://imagej.273.s1.nabble.com/Autocorrelation-using-Fourier-tp3686185p3686190.html

2010/12/14 Damon Poburko <[hidden email]>:
> Hello again,
>
>  Thanks for the tip. Padding the images up to the next power of two worked
> nicely. I didn't try smoothing the transition from the image to the padded
> region. Rather, I added a little bit of noise, then ran the correlation
> image through a band pass filter. Seems like it works just fine. Many
> thanks!!! I've been scheming for months over how to implement this, not
> knowing it was under my nose the whole time.


Dear Damon:

http://pacific.mpi-cbg.de/wiki/index.php/Stitching_2D/3D

It's even macro recordable. Uses a multithreaded FFT implementation,
and will work on 2D images and on 3D stacks. The padding is automatic
and done smartly with mirroring to prevent image edge effects.

Using it's internals, we put up a script to correct for 3D drift in 4D
image acquisitions. See "Plugins > Registration > Correct 3D drift"

Best,

Albert
--
http://albert.rierol.net