I just spent 8 hours getting a voice cloning program working. It's slow and clunky and missing some real basic features, but it's still super interesting.
Is Fakeapp still circulating? Is it any good?
cool painting site anon, nice find.
This is sweet
Some bad ideas
Use this https://github.com/CorentinJ/Real-Time-Voice-Cloning
to clone the voice of https://torrentz2.eu/18577d1d4fe1afe195d8350ab482d7b05567a0af
It would be really based
You can give it hits by putting little bits of color down.
Here is another bad idea: Facial Landmark/"Mapping" based (((Coincidence Detector)))
Assuming we are using DLib's Facial Landmark algorithm, and assuming (((they))) haven't done complete plastic surgery, we should be able to find (((them))).
Something like http://dlib.net/imaging.html#get_frontal_face_detector and http://dlib.net/imaging.html#extract_highdim_face_lbp_descriptors already exists.
Some primer: https://archive.li/XNuuN https://archive.is/KPZCB
The idea is through their eyes, nose, lips and chin we can deduce ethnicity.
thispersondoesnotexist would really come in handy if it could generate multiple pictures of the same "person" for use with various online identities
Ironically it is very easy for people to detect that the image is fishy as fuck
These two might help you in the long run
They would be passable for a profile pic or a thumbnail but looking closely they are obvious fakes.
I bet somebody has made a network of fake faceberg/twatter accounts with the details and pictures.