Jennifer Lawrence and George Clooney as you've never seen them before: Facebook creates software that distorts people's faces in video & # 39; s to make them invisible to face recognition & # 39;
- Facebook has fooled face recognition in a video for the first time
- It was said that Footage shows the effects on celebrities, including Jennifer Lawrence
- The software can help protect users against & # 39; misleading & # 39; deepfakes
Facebook has developed software to make people invisible to face recognition technology.
The & # 39; de-identification & # 39; program is intended to protect people from videos with & # 39; deep & # 39; style in which their faces can be edited on other people's videos.
These convincing clips become so advanced that it can be difficult to see which video & # 39; s are real and which are fake.
And there are concerns that in the future people can make images of others who do or say things that they have never actually done.
But Facebook AI Research now says it has a way to mislead the artificial intelligence that is used to make these videos while the original video still remains lifelike.
The company revealed its invention in an article entitled Live face identification in video.
And images show that they have tested it on clips from celebrities such as Jennifer Lawrence, Willem Dafoe, Sarah Jessica Parker and George Clooney.
In the & # 39; de-ID & # 39; clips, the faces of the speakers look smoother and airbrushed but are still recognizable.
Facebook said the process could work on live images and still images as well as recorded videos.
Basically, it works in a similar way to face-swap apps, which are popular on smartphones, by swapping someone's face with his own face.
This means that a somewhat distorted, computer-generated face is created with images from the past and placed on top of the real face.
As a result, they still look like themselves to the viewer, but a computer cannot retrieve essential bits of information that it could from a normal video or photo.
The videos regenerate the face of a person making them look the same – although somewhat distorted – for a human viewer, but manipulated so that artificial intelligence cannot identify their face (Photo: Jennifer Lawrence)
In their article, researchers Oran Gafni, Lior Wolf and Yaniv Taigman said: & # 39; Because facial technology is both useful and impactful, it also raises many ethical concerns.
& # 39; Face recognition can lead to a loss of privacy and face replacement technology can be misused to make misleading videos.
& # 39; In this work, we focus on video-de-identification, a video filtering application that both requires a technological leap over the current state of the art and that is benign in nature.
The software works in the same way as face-swap apps, which are popular on smartphones, by swapping someone's face with his own face (Photo: George Clooney)
The team said their software had defeated state-of-the-art face recognition and was the first to do it on a video (Photo: Willem Dafoe)
& # 39; This application requires the making of a video of a similar-looking person so that the perceived identity is changed.
& # 39; Allows the user to leave a natural-looking video message on a public forum in an anonymous way, thus preventing facial recognition technology from recognizing it. & # 39;
The team said their software had reported state-of-the-art face recognition and was the first to do this on a video.
She also boasted that it was able to preserve the person's original expressions and poses and worked on a wide range of ethnic groups and ages and on both sexes.
Facebook told technical website VentureBeat that it had no plans to roll out the software on a public platform.
The AI Research team will present their research next week at the International Conference on Computer Vision (ICCV) in Seoul, South Korea.
HOW DOES FACE RECOGNITION TECHNOLOGY WORK?
Face recognition software works by matching real-time images with a previous photo of a person.
Each face has around 80 unique nodes in the eyes, nose, cheeks and mouth that distinguish one person from the other.
A digital video camera measures the distance between different points on the human face, such as the width of the nose, depth of the eye sockets, distance between the eyes and the shape of the jaw line.
Another smart surveillance system (photo) can scan 2 billion faces in seconds, has been revealed in China. The system connects to millions of CCTV cameras & uses artificial intelligence to select targets. The army is working on a similar version of this with AI to track people across the country
This produces a unique numeric code that can then be linked to a corresponding code obtained from a previous photo.
A face recognition system used by officials in China connects to millions of CCTV cameras & uses artificial intelligence to set targets.
Experts believe that facial recognition technology will quickly catch up with fingerprint technology as the most effective way to identify people.
. (TagsToTranslate) Dailymail (t) sciencetech