Transforming Video Content Into Another Video’s Style, Automatically
Researchers at Carnegie Mellon University have devised a way to automatically transform the content of one video into the style of another, making it possible to transfer the facial expressions of comedian John Oliver to those of a cartoon character, or to make a daffodil bloom in much the same way a hibiscus would.
Because the data-driven method does not require human intervention, it can rapidly transform large amounts of video, making it a boon to movie production. It can also be used to convert black-and-white films to color and to create content for virtual reality experiences.
“I think there are a lot of stories to be told,” said Aayush Bansal, a Ph.D. student in CMU’s Robotics Institute. Film production was his primary motivation in helping devise the method, he explained, enabling movies to be produced more quickly and cheaply. “It’s a tool for the artist that gives them an initial model that they can then improve,” he added.
The technology also has the potential to be used for so-called “deep fakes,” videos in which a person’s image is inserted without permission, making it appear that the person has done or said things that are out of character, Bansal acknowledged.
“It was an eye opener to all of us in the field that such fakes would be created and have such an impact,” he said. “Finding ways to detect them will be important moving forward.”
Transferring content from one video to the style of another relies on artificial intelligence. In particular, a class of algorithms called generative adversarial networks (GANs) have made it easier for computers to understand how to apply the style of one image to another, particularly when they have not been carefully matched.
In a GAN, two models are created: a discriminator that learns to detect what is consistent with the style of one image or video, and a generator that learns how to create images or videos that match a certain style. When the two work competitively — the generator trying to trick the discriminator and the discriminator scoring the effectiveness of the generator — the system eventually learns how content can be transformed into a certain style.
A variant, called cycle-GAN, completes the loop, much like translating English speech into Spanish and then the Spanish back into English and then evaluating whether the twice-translated speech still makes sense. Using cycle-GAN to analyze the spatial characteristics of images has proven effective in transforming one image into the style of another.
That spatial method still leaves something to be desired for video, with unwanted artifacts and imperfections cropping up in the full cycle of translations. To mitigate the problem, the researchers developed a technique, called Recycle-GAN, that incorporates not only spatial, but temporal information. This additional information, accounting for changes over time, further constrains the process and produces better results.
The researchers showed that Recycle-GAN can be used to transform video of Oliver into what appears to be fellow comedian Stephen Colbert and back into Oliver. Or video of John Oliver’s face can be transformed into a cartoon character. Recycle-GAN allows not only facial expressions to be copied, but also the movements and cadence of the performance.
The effects aren’t limited to faces, or even bodies. The researchers demonstrated that video of a blooming flower can be used to manipulate the image of other types of flowers. Or clouds that are crossing the sky rapidly on a windy day can be slowed to give the appearance of calmer weather.
Such effects might be useful in developing self-driving cars that can navigate at night or in bad weather, Bansal said. Obtaining video of night scenes or stormy weather in which objects can be identified and labeled can be difficult, he explained. Recycle-GAN, on the other hand, can transform easily obtained and labeled daytime scenes into nighttime or stormy scenes, providing images that can be used to train cars to operate in those conditions.
Learn more: Beyond Deep Fakes
The Latest on: Deep fakes
via Google News
The Latest on: Deep fakes
- What 'deepfakes' are and how they may be dangerouson October 13, 2019 at 9:05 pm
The technology used to create such digital content has quickly become accessible to the masses, and they are called "deepfakes." Deepfakes refer to manipulated videos, or other digital representations ...
- California bans 'deep fakes' video, audio close to electionson October 12, 2019 at 11:40 am
SACRAMENTO, Calif. (AP) — California is trying to stop people from deceptively editing videos and audio in ways that are aimed at influencing elections. Legislation signed by Gov. Gavin Newsom on ...
- The biggest threat of deepfakes isn’t the deepfakes themselveson October 10, 2019 at 9:41 am
Subsequent forensic analysis never found anything altered or manipulated in the video. That didn’t matter. The mere idea of deepfakes had been enough to accelerate the unraveling of an already ...
- How deepfakes evolved so rapidly in just a few yearson October 8, 2019 at 9:00 am
For all the heightened concern over deepfakes being used to manipulate elections and sow chaos in countries around the world, the vast majority of them are actually much more juvenile in ...
- California takes on deepfakes in porn and politicson October 8, 2019 at 4:27 am
Sometimes, deepfakes are used purely for satire, such as to poke fun at government officials and the current political landscape. However, AII-manipulated content can take a dark turn when used to ...
- Porn, Politics Are Key Targets In 'Deepfakes': Studyon October 7, 2019 at 9:27 am
A report by the security firm Deeptrace identified 96 percent of deepfakes as "nonconsensual pornography," which use the images of female celebrities or others in videos manipulated using artificial ...
- Forget fake news—nearly all deepfakes are being made for pornon October 7, 2019 at 8:53 am
The internet is home to at least 14,678 deepfakes, according to a new report by DeepTrace, a company that builds tools to spot synthetic media. But most of them weren’t created to mess with elections.
- Deepfakes: For now women are the main victim, not democracyon October 7, 2019 at 6:09 am
The problem is that the apps can be used to harass and intimidate women. Although there's the obvious threat that deepfakes pose to democratic processes, the tech is also being applied in unexpected ...
- The Morning After: California's crackdown on political and porn deepfakeson October 7, 2019 at 4:37 am
Welcome to your week! Over the weekend, the state of California cracked down on deepfakes, New York's East River has floating LEDs to show if it's safe to swim and a torrent of Star Trek shows get ...
via Bing News