Jennifer Lawrence and George Clooney as you have NEVER seen them before: Facebook creates software that distorts people’s faces in videos to ‘make them invisible to facial recognition’
- Facebook has successfully fooled facial recognition in a video for the first time
- It said it Footage shows the effects on celebrities including Jennifer Lawrence
- The software could help protect users against ‘misleading’ deepfakes
Facebook has developed software to make people invisible to facial recognition technology.
Its ‘de-identification’ program is intended to protect people from ‘deepfake’ style videos in which their faces can be edited onto videos of other people.
These convincing clips are becoming so advanced it can be difficult to tell which videos are real and which ones are fake.
And there are concerns that, in future, people will be able to make footage of others doing or saying things that they never actually did.
But Facebook AI Research now says it has a way of fooling the artificial intelligence used to make these videos while still keeping the original video lifelike.
The company revealed its invention in a paper entitled Live Face De-Identification in Video.
And footage shows they have tested it on clips of celebs including Jennifer Lawrence, Willem Dafoe, Sarah Jessica Parker and George Clooney.
In the ‘de-ID’ clips the speakers’ faces look smoother and airbrushed but are still recognisable.
Facebook said the process could work on live footage and still images as well as recorded videos.
In essence, it works in a similar way to face-swap apps, which are popular on smartphones, by swapping someone’s face with their own face.
This means a slightly distorted computer-generated face is created using past images of them and put on top of their real one.
As a result, they still look like themselves to the viewer but a computer cannot pick up essential bits of information which it could from a normal video or photo.
The videos regenerate a person’s face which makes them look the same – although slightly distorted – to a human viewer, but manipulated so artificial intelligence can’t identify their face (Pictured: Jennifer Lawrence)
In their paper the researchers Oran Gafni, Lior Wolf and Yaniv Taigman said: ‘Since face technology is both useful and impactful, it also raises many ethical concerns.
‘Face recognition can lead to loss of privacy and face replacement technology may be misused to create misleading videos.
‘In this work, we focus on video de-identification, which is a video filtering application that both requires a technological leap over the current state-of-the-art, and is benign in nature.
The software works in a similar way to face-swap apps, which are popular on smartphones, by swapping someone’s face with their own face (Pictured: George Clooney)
The team said their software had beaten state-of-the-art facial recognition and was the first one to have done so on a video (Pictured: Willem Dafoe)
‘This application requires the creation of a video of a similar looking person, such that the perceived identity is changed.
‘This allows, for example, the user to leave a natural-looking video message in a public forum in an anonymous way, that would presumably prevent face recognition technology from recognizing them.’
The team said their software had beaten state-of-the-art facial recognition and was the first one to have done so on a video.
They also boasted it was able to preserve the person’s original expressions and poses and worked on a diverse range of ethnicities and ages and on both sexes.
Facebook told tech website VentureBeat that it had no plans to roll out the software on any public platform.
The AI Research team will present their research at the International Conference on Computer Vision (ICCV) in Seoul, South Korea, next week.
HOW DOES FACIAL RECOGNITION TECHNOLOGY WORK?
Facial recognition software works by matching real time images to a previous photograph of a person.
Each face has approximately 80 unique nodal points across the eyes, nose, cheeks and mouth which distinguish one person from another.
A digital video camera measures the distance between various points on the human face, such as the width of the nose, depth of the eye sockets, distance between the eyes and shape of the jawline.
A different smart surveillance system (pictured) can scan 2 billion faces within seconds has been revealed in China. The system connects to millions of CCTV cameras and uses artificial intelligence to pick out targets. The military is working on applying a similar version of this with AI to track people across the country
This produces a unique numerical code that can then be linked with a matching code gleaned from a previous photograph.
A facial recognition system used by officials in China connects to millions of CCTV cameras and uses artificial intelligence to pick out targets.
Experts believe that facial recognition technology will soon overtake fingerprint technology as the most effective way to identify people.
