I'm intrigued by the EditDistance idea. Right away I started trying to think of how I could implement it. It seems like in order for this to work, the edit dist

I'm intrigued by the EditDistance idea. Right away I started trying to think of how I could implement it. It seems like in order for this to work, the edit distance would have to be based on something more averaged than a single pixel color due to the various types of image compression (jpeg being particularly troublesome for this task, I would imagine). Perhaps the image can be split up into four-pixel blocks whose constituent parts can all be averaged together to form a new color which can then be used as part of the matching. I must admit however that I have doubts that an edit distance for images would yield interesting results unless the pixels were split into a grid, or maybe something like a quad-tree. Then the edit distance could be applied to each portion of the grid to see how many adjacent sections of the grid are also found to be adjacent in other images.