Potential contribution: subpixel cross correlation

[Stéfan van der Walt]

On Sun, Apr 27, 2014 at 2:48 AM, Michael Sarahan <msarahan at gmail.com> wrote:
> Your example highlights a critical issue: handling color data.  I've only
> ever played with grayscale data (electron microscope images).  I tested my
> code with your example image, and it falls over.  Do you have any tips for
> what special handling I might need to do (convolutions, finding maxima)?  I
> can just grayscale the input images, but that doesn't feel like an optimal
> solution.

Perhaps Juan can answer this one--he is our most vocal 3D proponent!

In the meanwhile, getting a version up and running with tests in 2D is
a great start.

Stéfan