Okay, I see your point. I find this a bit suspicious though. The process in the link you gave (and in in some others I found) shows how images can be manipulated in order to remove the fisheye effect after the image has already been taken. In principle this is fine, but for a standard of evidence I find it weak. This is because there is no direct complement to the specific lens effect. What I mean is that the original pictoral information has been lost. Yes, you can later attempt to reverse the distortion effect, but the original information is lost. A good analogy is this: you have a bucket with a few grains of sand in it (the original information) and then fill the whole bucket with sand (the lens distortion), you then take out most of the sand, but since you don't know how much sand you had originally, you make estimate (the reversing process). Thus you are left with some amount of sand that looks right, but is not really identical to the original. I hope I made my point clear.
