[ImminentFate]

On your site I can consistently get 100% by looking at the backgrounds since they generate in somewhat inorganic patterns.

[nuccy]

Also after watching the video from the StyleGAN2 team https://drive.google.com/file/d/1f_gbKW6FUUHKkUxciJ_lQx29mCq... now I know that original StyleGAN, images from which are apparently used for this "game", produces faces with a "water droplet" and phase artifacts, so I was able to spot few fakes just by looking for those things.

[Veedrac]

Only whichfaceisreal.com uses the original StyleGAN. The github.io links use StyleGAN2.

[ebg13]

That's speculative at best. What's inorganic about the background in https://veedrac.github.io/stylegan2-real-or-fake/Curated/ffh... ?

[gajomi]

How is it speculative? The background does not always allow for the discrimination to be made but in the random sample of ~20 faces I looked it was the main factor in maybe 1/4 of the cases (I was %100 accurate on this random sample). Of course, my random sample is not your random sample. We could probably do a controlled experiment to get at these kinds of attributions systematically.

I might also say that a second major discriminating factor is skin texture, especially at boundaries.

[TomVDB]

When not the ears, it's the background.

If the real pictures had the background removed, I'd have a very hard time scoring 100%.

[Veedrac]

That's the point of the hard version.