oldtigger wrote:
dreams are always a bit fuzzy in hindsight but i got the idea they generated the 10 images using variable edge detect and the image you chose was actually stacked using contrast values so that only the edges or transition points of the original were effected, not the whole image.
It seems that could be accomplished easily in software without requiring any other images.
kind of like an blurred high-pass overlay... Am thinking that edge detection algorithms combined with sharpening would do some of that, but I'm sure the magic is in selecting those areas you want to enhance...