Doesn't matter. One fingerprint does not correlate with one valid image anyway, and not with
two either. I don't know what the correlation is, but I would bet on thousands of valid images
The problem is that a fingerprint isn't even presumed to be unique for every living human
(Wikipedia has three examples of false positives, from Europe and the US from 1997 to date,
that is clearly a limited population, time and fingerprint database). And while the
fingerprint has molecular resolution, an image of one hasn't - so uniqueness definitely drops
quit a bit. So instead of one in 100 million, you may very well drop to one in 100 thousand.
The pigeon hole principle certainly applies.