Posted on 12/18/2018 3:10:11 AM PST by Freeport
Believe it or not, all these faces are fake. They have been synthesized by Nvidia’s new AI algorithm, a generative adversarial network capable of automagically creating humans, cats, and even cars.
The technology works so well that we can expect synthetic image search engines soon — just like Google’s, but generating new fake images on the fly that look real. Yes, you know where that is going — and sure, it can be a lot of fun, but also scary. Check out the video. It truly defies belief:
According to Nvidia, its GAN is built around a concept called “style transfer.” Rather than trying to copy and paste elements of different faces into a frankenperson, the system analyzes three basic styles — coarse, middle, and fine styles — and merges them transparently into something completely new.
Coarse styles include parameters such as pose, the face’s shape, or the hair style. Middle styles include facial features, like the shape of the nose, cheeks, or mouth. Finally, fine styles affect the color of the face’s features like skin and hair.
According to the scientists, the generator is “capable of separating inconsequential variation from high-level attributes” too, in order to eliminate noise that is irrelevant for the new synthetic face.
For example, it can distinguish a hairdo from the actual hair, eliminating the former while applying the latter to the final photo. It can also specify the strength of how styles are applied to obtain more or less subtle effects.
(Excerpt) Read more at tomsguide.com ...
It’s also possible to simulate (not create) images of real individuals doing and saying things that the real counterpart never did or said.
The green chick went on to play Barbara Gordon, alias Batgirl, as the Batman TV series was fading. She was also a major hottie in “Kissing Cousins” with Elvis Presley.
“endless retakes”
Even better, I can do endless retakes if needed as times change. “Long Tail” characters that can be updated.
I could see them being used to recreate concert experiences, imagine being able to see a reasonable facsimile of a Beatles concert, with figures that look and move just like the real thing.
“If the b____ is green, there must be something wrong with the p____y!” - Eddie Murphy
That’s actually a pretty neat idea - I might even pay money to see a high-quality Keith Moon or Jimi Hendrix. It’s not too hard to imagine a time in which you could purchase a system that allows you to assemble your own super group of vintage rock stars (I think I’ll call mine “The Dead 27’s”).
All the Democrat voters you could ever want - now with photo IDs! What’s the matter with you Republicans who claim vote fraud with no evidence? /s
Peter Cushing acting in a movie made in 2016? Tom Hanks shaking hands with John F. Kennedy?
Donald Trump shaking hands with Hitler.
If Susan Oliver was a knockout then Yvonne Craig was a Five Point Palm Exploding Heart Punch.
Would still not be legal in the US.
Section 2256 of Title 18, United States Code, "defines child pornography as any visual depiction of sexually explicit conduct involving a minor (someone under 18 years of age). Visual depictions include photographs, videos, digital or computer generated images indistinguishable from an actual minor, and images created, adapted, or modified, but appear to depict an identifiable, actual minor. "
So even if no real-life child was involved, it would still be illegal to depict.
I played with Blender some ages ago.
If commercially-available software is this good, I would guess that there are classified capabilities which are even better.
The CIA or DOJ could frame anyone they wanted.
If I understand this correctly, this means that the DA would have to have an actual child that they could produce that looks like the computer generated child.
Maybe they could do it maybe not.
Sorry fake faces using shading on statistically boiled down REAL FACE DATA isn’t that big an achievement.
One - can it do it in real time (and in 3D)
TWO - can it ANIMATE those faces in changing environment (the muscle and skin/hair movements are far more complex)
THREE - those faces are mostly small pictures and detail needs to work in REAL closeups
FOUR - they also look rather polished - various realistic defects that real faces have are a whole nuther massive amount of data that would have to be boiled down (and categorized and attributized) to be truely ‘random’
I guess it’s like a Mary Anne vs. Ginger thing, except they’re both Mary Anne.
...couple this with speech synthesis, where text-to-speech capability includes spoofing a specific persons voice patterns and you can fake anything about anyone.
https://www.csoonline.com/article/3196820/security/vocal-theft-on-the-horizon.html
Inevitable? Yes.
Scary? I think not.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.