Nvidia (a client of the author) has done a lot of exciting things lately, from creating workstations designed to design the metaverse, to digital assistants that evolve into human digital twins, to tools that could allow anyone to create compelling art. One of the most interesting tools is Generator StyleGAN, which creates people's faces by mixing images.
The training set for this AI-based offering contains 70 high-quality PNG images (each with a resolution of 000 x 1024 pixels) that allow the user almost unlimited flexibility of source material.
StyleGAN has been around since 2018, became more widely available in 2019 when the source code became open source, and is now in its third permutation. StyleGAN3 was released last October.
Benefits for those of us who work with images include the potential ability to create images from large pools of copyrighted source images without facing copyright issues or worrying about copyright infringement. And as the process evolves to include other images (it's basically an image fusion engine), it could allow you to blend professional photos from various sources to create unique and beautiful images or paintings created from memory or imagination with little or no effort. no connection to anything real.
An AI-based image blending tool like StyleGAN could radically change and improve a number of industries and practices (or be used for more nefarious "deep fakes"). We are going to explore.
Automated crime cartoonists?
I see many criminal proceedings on television; there is usually a segment where someone sits down in front of a cartoonist to create an image of a criminal that he has been watching. This entire process could be automated by a conversational AI. The witness may be shown an evolving image with sample features that are shuffled on demand until the image matches the victim's memory. The end result would be a photorealistic image that could be used by facial recognition programs to quickly locate the offender. (The collateral damage would be that there would be no need for law enforcement cartoonists.)
One area where this technology could have a big impact is locating kidnapped children. The AI could rapidly age the image of the child so that it is better identified later in life.
Marketing, TV and Film
Many marketing materials use stock images or production models. The problem with the former is that those same images can be used in other campaigns, inadvertently linking disparate campaigns together. For example, if the same image is used in an ad for a drug and a restaurant, customers may associate the two and avoid the restaurant. The same problem could result from using a live model that then ends up in another campaign, as some actors and models move between contestants. And living models/actors can have personal issues that can hurt a brand or ad campaign.
But using combined images and videos from something like StyleGAN means you can create an image that may be copyrighted to your company, unique from any stock image, and not tied to any actor or model, living or dead. The result is lower cost and, more importantly, lower risk. You will get results faster and the need for models and actors would be reduced. You can only use actors in 3D image suits that hide their identity, and with advances in metaverse tools and 3D image players, you may not even need them. It also brings us closer to not needing actors for movies.
Human digital twins?
Another area Nvidia is exploring involves creating digital twins for the metaverse. And as the AI behind these twins improves, they will become more indistinguishable from the source material. When this happens, who owns the result? You can argue that an employee should own their digital twin. But if a tool like StyleGAN is used to mix up an employee's images and skills, that position becomes more precarious; a company might be able to defend its property from the result. (I hope future employees and unions have major issues with something like this being used to displace employees without compensation.
a mixed future
The ability to remix source material that may (or may not) be copyrighted on a massive scale is compelling, especially if it removes potential legal issues. Nvidia's process uses a verified image source which eliminates legal exposure, but tools like this shouldn't rely solely on photo databases; could be used in images of public figures taken from social media posts, movies, or other publicity materials.
At some point, I hope this technology will force a rewrite of copyright laws dealing with composite images. At the same time, they would reduce the amount of effort and cost required to create photorealistic images and movies that can be used in business and entertainment. It is an early example of major changes in current business practices and associated income for those who work as models, actors or directors, and for artists tasked with creating images that define remembered events.
Tools like StyleGAN will redefine the future of virtual media for business, government, and entertainment.
Copyright © 2022 IDG Communications, Inc.