Slashdot Mirror
'This Person Does Not Exist' Website Uses AI To Create Realistic Yet Horrifying Faces (inverse.com)
A website that uses AI -- Generative Adversarial Networks (GANs) -- to generate photos of people who do not exist is circulating on social media and forums this week. A news writeup adds: Every time the site is refreshed, a shockingly realistic -- but totally fake --picture of a person's face appears. Uber software engineer Phillip Wang created the page to demonstrate what GANs are capable of, and then posted it to the public Facebook group "Artificial Intelligence & Deep Learning" on Tuesday. The underlying code that made this possible, titled StyleGAN, was written by Nvidia and featured in a paper that has yet to be peer-reviewed. This exact type of neural network has the potential to revolutionize video game and 3D-modeling technology, but, as with almost any kind of technology, it could also be used for more sinister purposes.
I pulled up the website and the picture looked a lot like me.....
The only thing necessary for evil to triumph is for it to be pitted against a slightly greater evil
Bet you've not heard that term in a while!
The site for me is loading the image slower than an 80's fax-modem set to highest resolution.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
A friend who lives in a repressive country where over 90% of the media is pro-government propaganda told me that the when he does a reverse image-search of the little "author image" next to the opinion pieces lauding the government's actions, neither Google nor Tineye can find the person online. It seems that the opinion pieces are authored by supposed "journalists" who's face can only be found next to the opinion piece - and nowhere else online. The authors names and their face images are fictional, even though their faces appear to belong to a real person. It seems like this tech was around before Nvidia supposedly "pioneered" it.
Why did the chicken cross the road? Because Elon Musk put an AI chip in its head.
First pic was a quite attractive woman, not much luck with the next 5.
Also the Slashdot effect seems to be in progress so its pretty slow ...
I can remember dabbling with fractals as a basis for creating artificial landscapes back in the 80's. (And no, it wasn't my lawn which you can kindly depart from.)
It's interesting to see the level of detail, and the types of asymmetries and other 'imperfections' . It'd also be interesting to determine statistical probabilities of a reasonably close match to an actual person.
After the first 5 or 6 images, I kept getting the same image of a black woman with strange eyes.
Update:
403 Forbidden
nginx/1.14.0 (Ubuntu)
I think we just slashdotted them ...
Slashdotted again!
...of downloading porn over dialup.
Thank Jeeberz for broadband!
It's not 100% realistic yet. The faces look good, but with the hair and body there are a lot of issues. I pulled up one picture and the face is a significantly lighter tone than the neck/chest area which leads to issue number 2, the woman does not appear to have a neck. The ears are just close enough to realistic to notice they aren't right, and the hair coming down over the body isn't correct either. But the facial features are really good.
The only thing necessary for evil to triumph is for it to be pitted against a slightly greater evil
Seems we’ve done it! Did they run it on a smart card or what?
Yes i would smash it indeed
Someone should write a bot to post each new photo to Facebook (along with a randomly generated name) to salt their facial recognition algorithm.
Support Right To Repair Legislation.
great site!
Looks as if it just models the face in 3d, then applies a texture from another photo (or several parts). I've seen exactly the same pose/size face three times now but with different people
So it's just a fancy 'magic wand'
Did the headline writer read the word "shocking" in "shockingly realistic" and decide the website - which he didn't bother going to - must be about something scary ?
Or is he just an idiot ?
A purple blob along the nose attached to the eye.
Having cycled through some more, my real question is: Whose stock images are they using to generate these fake ones?
I assume they have a huge database of faces and my question is if they are all stock footage with a model release, or if they are just pillaging social media sites and not letting anyone know they've been using their likenesses for transformative purposes (which is all this sort of GAN modelling is, transformative works based upon a large corpus of publically available but not commercially licensed media.)
I was pretty sure doing so was considered copyright infringement, but if things have changed I hope someone can update me on those legal details.
Some of the pics have noticeable artifacts that give the appearance of a severe scar... or possibly gills in some cases.
Sig. Sig. Sputnik
didn't see any horrifying faces just normal looking ones
This is *exactly* the sort of disinformation I'd like to promulgate for Faceybooks facial image recognition !
As it is, the created faces just look like random people to me. There is nothing "horrifying" about it.
Hey wait! That's me! Does that mean I don't --
Generate the rest of their bodies as well and we'll finally have a replacement for Tumblr!
I am not interested in articles about life extension advancements.
I didn't know the industry needed up'ending, but this could forever change "Big Hero-shot".
It's a bunch of unremarkable human faces. Where's the horror described by the breathless headline?
These losers can use this site to generate a picture of their "girlfriend" and post it to sites like Facebook and no one will be the wiser. Creepy, but not as creepy as doing the same with the photo of a random (but real) girl they found on the Internet.
I would say but /. is supposed to be safe for work.
It's obvious the neural network was trained on stock images, including lots of celebrities. I saw Angelina Jolie's eyes, Brad Pitt's jaw, Caitlin Jenner's hair.
...laura
They all have nearly perfect teeth. And the shape of the two front teeth is really similar between all people that show their teeth. I don't think these are real people...
What if you see one of the images and fall in love with the person in it? They don't exist, but you can't get the idea of them out of your mind and you are ruined. Forever.
That would be a nightmare.
"Wow from stratch! AI can generate these faces!" Move around a couple of pixels change a few and say wow our technology can generate these faces from scratch. Humans give too much credit to the human race. Especially those who wield the american flag with pride, we are the most best good people in the universe, we are the pinnacle of truth.
1991 called and they want credit back for this new face "morphing" thing.
http://www.criticalcommons.org...
Pause any morph in the state between two faces and you have the very realistic looking face of a person who does not exist. Couple that with our modern software algorithms that identify various landmark points on a person's face (used by SnapChat and other filters to put sunglasses on a person's face in the right location, etc), and voila... you have infinite face creation with no manual input by blending input images together. It's very clear in some of the images on the website from TFA that they are using actual face photos as their source input, so what they are doing is a derivative on the face morphing, perhaps using multiple inputs instead of two.
Better known as 318230.
Finally! A new picture source for my online dating profile!
Hmmm... I loaded the first page and all I got was a picture of me. What's that about?
Some of those faces can be instantly identified as fakes. Some have eyes that are beyond uncanny valley and down right hideous.
Others... well I saw at least 2 images where I couldn't find a fault despite looking for many minutes.
They can make all the fake profiles they want. reference
Yes, we would all be repelled by that image. A portrait of your face could have the same effect as Goatse, Tubgirl or Lemonparty.
That is a bunch of face parts stitched together. And you can see the stitching. I'm not impressed. A man's mouth stapled to a womans face might fool AI but it doesn't fool me.
That book with four separate sets of pages, so you could choose the hair, eyes, nose and mouth to make hilarious combinations.
Good times.
"Nine times out of ten, starting a fire is not the best way to solve the problem." - my wife
Those are good but still look somewhat uncanny, like if all the pieces are glue together.
What do they do, just re-arrange the eyes, nose, and mouth on a perfectly good picture?
I looked at some of the pics. Either the pictures are real, or they're composites of real people. There's too much background detail. The reflections in eyeglasses show content. Almost every pic had some easily identifiable artifact like a smudge, a paperclip growing out of the top of someone's head, or the temple from a pair of eyeglasses just hanging out without the rest of the glasses.
For me to buy into this, just get the semi-fake person to start sending tweets with the AI-driven fake tweetbot.
The Russians have won. They have made the world a cesspool of distrust, greed, fear and hate.
Let's flood social media with millions of these so you can no longer tell real users from fake ones. Here's your big data. Real big. The biggest. Go forth and market to no one.
Or are they manipulated images of real people ?
There were several of the women I'd like to see more of, if you know what I mean.
Strictly for research purposes of course!
Tens of refreshes and no elderly faces rendered
Do seniors exist in the AI world?
Is it just me, or do a lot of the images suffer from heterochromia?
Also, anyone with glasses seems to have problems.
So not only are they realistic, but unique, almost vaguely like a fake name generator that also creates a documents and credentials paper-trail. When deep fake bots like this can reliably pass the Turing test and generate nonlinear situations such as talking to a human over Skype, and then appearing to take their laptop outside or around a house, that's also deep fake generated, it will become increasingly more difficult to interact with people online, especially when the sales calls become 95% posing as human salespersons.
The real magic isnt that it is able to create new faces, but in fact, how it is able to more accurately blend the individual real-world body parts (ie, the eyes, hair, mouth, wrinkles all come from real-world examples) in the scene by understanding the correct lighting and textures. We have been able to randomly generate non-existent faces for decades, but doing so, so the parts blend well has always been the hard part. It's a Face-Time filter applying more passes. I still fail to understand why the buzzword: AI needs to be linked to this besides as a way to garner more interest and funding.
You'd expect a real journalist to publish more than one story in their lifetime and to use the same image when publishing in the same paper/site
Most of those are probably they same few of you who seem to have an unhealthy obsession with the guy. Seek professional help.
Wanna buy a shirt?
https://www.redbubble.com/people/stealthfinger/shop?asc=u
It seems to create weird artifacts around the edges of hair and ears.
No sig. Move along - nothing to see here.
I got a better name for the service. I guess you can't trust anybody on the Internet (anymore).
"Everybody's naked underneath" -- The Doctor
Indeed, this is sad isn't it?
The only views he gets come from people laughing at him. We left several comments on his channel and he deleted them all.
If you look here:
https://socialblade.com/youtub...
You will see that the comment count is higher than what's actually on his channel.
When he first launched his channel, we even ran click bots against his channel to artificially inflate his views. We told him about it but he wouldn't believe us until we told him in advance that we would target specific videos with no views.
Before he realized that we were really the ones inflating his views, he regularly came to Slashdot and bragged about how successful he was!
Sad, so sad...
CROFLOL! True enough, while running the clickbots, he also came here all the time to spam his video links because he thought it was efficient. But it was clickbots views, not his spam being efficient.
CROFLOLOLOLOLOLO!
LOL! and, of course, the click bots didn't follow his Amazon spam links on youtube and didn't make him a penny!
But back then, he said success was coming and that soon, creimy the mountain would be rich!
What a brilliant long tail revenue stream!
Sad, so sad...