Facepalm As Researchers Crack Advanced Facial Recognition Tech With Simple Makeup
The researchers at Ben-Gurion University explained that facial recognition is widely used in subways, airports, and workplaces to automatically identify individuals. In this experiment, the ArcFace face recognition model was used with 20 blacklisted participants who would be flagged in a real-world facial recognition setup, including different angles, lighting, and cameras.
Then, digital makeup was applied to images of participants, which were then used to test the facial recognition system virtually. To assess this in the real world, a makeup artist also applied random and adversarial makeup to participant’s faces as it was in the digital image. When testing the digital images, the researchers found that the system was unable to recognize all the participants, whereas the physical makeup changes made it so the system could only spot 1.22% of the video frames that had a blacklisted participant.
However, it should also be considered that the system only spotted the blacklisted people 45.57% of the time without makeup and 33.73% of the time with natural-looking makeup. Despite this, the reduction in detection of the facial recognition system was still significant overall. Interestingly, it also appears that females using a makeup attack against facial recognition have slightly higher odds of success than men.
The concerning part of all of this is that the research also explained this “attack has a high level of inconspicuousness and thus can be launched in the real world.” As such, though it seems something straight out of a Bond or Bourne film, some cheap makeup may sneak you past facial recognition surveillance systems if you are running from the law.