Histogram matching

Histogram matching seems like a pretty good segmentation technique. The first row below is the left image, the right image, and the differences after an affine transformation. The second row is the same, but with histogram equalization. 

The average pixel difference dropped from 38 to 4, and I got it down to < 2 by cropping out the sections of the image that aren't visible in the other view.

It should be noted that there will still be some residual error from the affine transformation, so < 2 is pretty remarkable. I just hope this stands up to fluorescence!