It would be faster, but only a little. Instead of scanning 10000+ files to find a match, it would only need scan 10-20.
It would also prevent people from looking at all of the pictures by incrementing (or whatever) the image number.
Futher than that, I'm not sure.