You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I propose that we implement the following as a thunder registration method:
Manuel Guizar-Sicairos, Samuel T. Thurman, and James R. Fienup, "Efficient subpixel image registration algorithms," Opt. Lett. 33, 156-158 (2008). link
Essentially, images are transformed using an FFT, up-sampled to an arbitrary degree and aligned via cross correlation, the benefit of this algorithm is the efficiency of the up-sampling, and the fact that it is sub-pixel. It works quite well and porting the code should be straight forward.
The text was updated successfully, but these errors were encountered:
Thanks @andrewgiessel , I think this would be terrific to have! Take a look at the image registration API and maybe try putting together an implementation? Happy to jump into a discussion in the chat room if you get stuck. You can also check out this WIP PR from @poolio that adds another, different registration method (#100).
I propose that we implement the following as a thunder registration method:
Manuel Guizar-Sicairos, Samuel T. Thurman, and James R. Fienup, "Efficient subpixel image registration algorithms," Opt. Lett. 33, 156-158 (2008). link
MATLAB implementation
Original Python Port
My generalization for neuro data in d_code
Essentially, images are transformed using an FFT, up-sampled to an arbitrary degree and aligned via cross correlation, the benefit of this algorithm is the efficiency of the up-sampling, and the fact that it is sub-pixel. It works quite well and porting the code should be straight forward.
The text was updated successfully, but these errors were encountered: