This Person Does Not Exist Com

Generator of fake portraits

People tend not to recall about the issue that neural networks accept on our lives, because usually, we see the result of its piece of work and not the "confront" of a neural network. Perhaps that is why the generator of fake photos became the main topic of discussion for several weeks in the media devoted to applied science at the end of 2020. Not anybody was able to estimate that AI could generate a realistic confront of a not-existent person in a couple of seconds. Fake portraits look very realistic and it's frightening. If AI can create faces for itself and can text like real people, so what is going to happen side by side?

Generator of faux faces of non-existent humans

We are talking about the website thispersondoesnotexist.com ("this person does not be dot com") and are going to tell of the history and areas of application. The way the generator works volition exist explained further.

The AI face generator is powered by StyleGAN, a neural network from Nvidia developed in 2018. GAN consists of 2 competing neural networks, 1 generates something, and the second tries to observe whether results are existent or generated by the first. Training ends when the first neural network begins to constantly deceive the second.

An interesting point is that the creation of photographs of non-existent people was a by- product: the main goal was to train the AI to recognize fake faces and faces in general. The company needed this to improve the performance of its video cards past automatically recognizing faces and applying other rendering algorithms to them. However, since

the StyleGAN code is publicly available, an engineer at Uber was able to take it and create a random face generator that rocked the cyberspace.

Nigh the generator

For the user, everything works very simply. As soon as yous are on the website random face up is generated. Y'all can download the picture if you want. Refresh the page if you don't similar the person that yous are seeing. If you see the aforementioned face, simply wait a couple of seconds, and refresh the page again. The website shows the results of the generator's piece of work (which are updated every two-3 seconds) not the generator itself.

How to recognize an prototype of a false person

It is almost impossible to recognise an image of a false person. AI is and then adult that 90% of fakes are not recognized by an ordinary person and 50% are not recognized past an experienced lensman. There are no services for recognition. Occasionally, a neural network makes mistakes, which is why artifacts appear: an incorrectly aptitude pattern, a strange hair color, and and so on.

The only thing y'all demand to exercise is take a closer look: humans' visual processing systems are far stronger than computers', then it is possible to recognise forgery past detection.

Jevin West and Carl Bergstrom created a website called "Which Face Is Real", which is focused on teaching people to be more analytical of potentially simulated portraits. Before making suggestions that a person in a photo is existent, in that location are several things that need to be considered.

One of the near mutual ones is symmetrical problems, in detail eyeglasses and earrings.

Aforementioned uneven issues with teeth are quite common besides. Expect for odd characteristics similar pixels and repeated incisors. Fake hair, in general, may seem with some glow effectually it or wait likewise directly and streaked, once again, with visible asymmetry.

Study groundwork advisedly. If it's faux, it may include unusual distortions in shapes and lines, or have a torn appearance overall. Drain-through occurs in bright colors that stand out from the background onto a imitation person's caput.

Some other distinguishing feature of the NVIDIA "StyleGAN algorithm" is glossy "water splotches".