Are you looking for unique face photos? Generated Photos offers you an exceptional portrait gallery. This site features photos of people who are not. The American start-up produces fake faces (Fakefaces) hyper-realistic thanks to artificial intelligence. These creations foreshadow the development of generative media.
Generated photos: new method of personalization?
The models in these photos do not ask you for royalties. They also don’t get paid for using their image. And with good reason: they don’t exist. Indeed, the start-up artificially created thousands of fake faces and put them on their website. Simply select the face that suits you based on gender, age, type, eye color, hair color and even an expressed emotion.
Like free image banks, these faces are also free to use for personal use as long as you link to the site. Generated photos. The site also offers paid licenses for commercial use of the photos.
This new type of image bank is the result of the 100K Faces project: a database created by taking more than 29,000 photos of 69 different models over the past two years at the team’s studio. The start-up then used a machine learning process to put these together new faces bigger than life.
To read >>3 free and license-free image databases
With generated photos, the conventional beauty changes
We see the advantages of this in communication and marketing artificial faces to identify a chatbot, create personas, or personalize content:
- A very large and precise selection of photo models
- completely royalty free images
ces facefakes Complete the range of deepfakes used in communication. Tyler Lastovich, engineer and entrepreneur behind the project, even sees it as a new paradigm: “Conventional beauty is changing. Our systems don’t understand or care whether a created person is “beautiful” or “ugly.” They care whether a person looks human or not. With generative media, it is possible to construct any combination of physical attributes. We will no longer have to rely on a few supermodels to set global standards.”
The arrival of generative media
The term “Generative Media” is born
A generative medium is a digital process that enables the creation or distribution of algorithm-generated content, information, or data.
Generative media is already present in the field of art (musical or visual works produced by artificial intelligence have become commonplace) and is now also influencing marketing and communication.
The technology used is called Generative Adversarial Network (GAN). Specifically, the machine learning algorithm used by Generated Photos is called StyleGAN (an NVIDIA resource).
Put simply, a GAN is two neural networks working against each other. The first network generates a fictional face from raw data; The second network verifies whether the statement is correct or not by comparing it to a database of real images. The first network is informed about the correctness of the result. And this “trial and error” learning process is repeated until realistic results are achieved.
The principle of GAN technology
Tyler Lastovich announces that this technology will not only be used to create “fakes”. She can also embody people from the real world in a television series, for example. One could imagine that your family members could thus play real roles in the next successful Netflix season.
Generative media is also paving the way for this an audiovisual production accessible to everyone. As Tyler Lastovich once again points out: “With generative media, a single creative person can create an entire movie just by sitting at the computer.”
That person doesn’t exist, and neither does his cat
The 100K Faces project had a few predecessors like https://www.thispersondoesnotexist.com. Nobody is real on this site. Just refresh the page to see a new face. A feline version of these artificial creatures is also available at https://thiscatdoesnotexist.com. The results are sometimes very surprising or monstrous, but are constantly improving over time as the algorithm is “trained”. This is also the problem of machine learning the development of which, with the advent of generative media, will occupy a prominent place in the field of communication.
Written by
Philippe Gerard
Philippe Gerard, Manager at Cegos, leads digital training and digital transformation projects. He is also a co-author of The Communication Toolbox for Dunod editions.
Learn more
