This AI System from NVIDIA Creates Realistic Looking Images of Artificial Faces
Researchers at NVIDIA have developed an adversarial neural network that produces highly detailed artificially generated images people’s faces. It is increasingly evident that we will not be able to believe most of what we see is real in the not-to-distant future.
NVIDIA has developed a system that employs a GAN to create far most realistic-looking images of people. While some of the images produced are obviously flawed, some are almost indistinguishable from the real thing. The rapid rise of this technology needs to stimulate discussion on not only the beneficial uses, but also on the potential unethical applications.
“We feel that convincing realism may now be within reach.”
In the paper, published online, the team showed how they added new layers dealing with higher-resolution details until the GAN finally generated images of “unprecedented quality.” Judging from the examples provided by NVIDIA, many viewers would have trouble knowing that some of the images were computer generated.
The system uses an algorithm called a generative adversarial network (GAN) to create the faces.
Artificial neural networks are systems developed to mimic the activity of neurons in the human brain. In a GAN, two neural networks vie against one another. One of the networks functions as a generative algorithm, while the other challenges the results of the first, playing an adversarial role.
The researchers conclude:
While the quality of our results is generally high compared to earlier work on GANs, and the training is stable in large resolutions, there is a long way to true photorealism. Semantic sensibility and understanding dataset-dependent constraints, such as certain objects being straight rather than curved, leaves a lot to be desired. There is also room for improvement in the micro-structure of the images. That said, we feel that convincing realism may now be within reach.
With obvious applications in entertainment, games and virtual reality, this work, and others like it also increasingly calling on us to question what we see. Along with images and video, other research is also underway that has greatly increased the believably of artificially generated speech. With artificial intelligence being used to puppeteer a model of a real or fake person, can we believe anything presented as fact on television or the Internet?
Judging by the present state of the news industry and politics, you might already be.