News for hackers. Stuff that matters.

A comment on Two recent studies on judging authentic vs. deepfake images. Parent comment

Re:Two recent studies on judging authentic vs. deepfake images by victor9000 on Thursday March 10, 2022 @11:32AM (#30628793)
The biggest problem for me is that the mouth movements seem completely off.
[ Parent | on HN ]
Re:Two recent studies on judging authentic vs. deepfake images by stavros on Thursday March 10, 2022 @01:19PM (#30630396)
I think that's because they used the original fake to make the deepfake. If they had used an actual actor (with a completely different face but accurate mouth movement) I think it would have been spot on.
[ Parent | on HN ]
Re:Two recent studies on judging authentic vs. deepfake images by SECProto on Thursday March 10, 2022 @12:55PM (#30630038)
It's off on both sides, but to me it looks more off on the deepfake side
[ Parent | on HN ]
Re:Two recent studies on judging authentic vs. deepfake images by adamsmith143 on Thursday March 10, 2022 @01:42PM (#30630706)
That's been an issue for all CGI, I guess our ability to spot even the slightest imperfection in how mouths look has us well stuck in the uncanny valley
[ Parent | on HN ]
Re:Two recent studies on judging authentic vs. deepfake images by mywacaday on Thursday March 10, 2022 @02:24PM (#30631274)
The firs comment on the video says it would be better if they had access to the original video and not just the original fake.
[ Parent | on HN ]
Re:Two recent studies on judging authentic vs. deepfake images by jagged-chisel on Thursday March 10, 2022 @12:19PM (#30629464)
His top lip doesn't move when he speaks
[ Parent | on HN ]