Liveness Detection: The Next Step into Facial Recognition
Face authentication was the middle child of biometrics for years, being unnoticed while standing aside to more popular biometric modals such as the fingerprint. It all changed in November 2017, when Apple announced face authentication as one of the main features in its flagship device, the iPhone X. Since then, the middle child of biometrics didn’t look so awkward anymore - even other tech juggernauts like Samsung started adopting the tech as well.
As we’ve all learned, with great power comes great responsibility. Shortly after the iPhone X was released, there were reports of hacks – the most notorious coming from the Chaos Computer Club in Germany, claiming they were able to bypass the facial biometric system in minutes. The dilemma quickly unfolded: can facial recognition systems be used as a valid security modal, or is it just trending new tech? The answer is yes and no – to both questions.
The biggest issue surrounding face authentication systems is presentation attacks, which occur when the attacker submits a biometric sample to fool the system into granting access or permissions. In face authentication, there are three main spoofing techniques:
- Print attacks: the attacker displays a picture, either print or digital, of the individual they are trying to impersonate.
- Video attacks: a close-up video of the victim’s face is used. Video attacks are a step up from print attacks since movement is added to the fake biometric sample.
- 3D masks: a 3D mask of the individual being embodied is worn. This approach is the most advanced since it has the capacity to fool depth sensors as well.
Even though middle-schoolers have spoofed sophisticated face authenticators with selfie cutouts (an Irish politician had his laptop “breached” by his son who used a campaign flyer’s picture), the biometrics industry isn’t giving up. The latest attempt in creating an anti-spoof face authentication system consists of adopting facial liveness. This technique entails gathering and analyzing data points from one’s face to determine whether it is a real, live person or a presentation attack. The first set of liveness techniques came in the form of response tests, where the user had to perform specific random actions such as smiling or blinking to the camera. Even though this added security layer was an improvement, it was not a spoof-proof or frictionless approach, diminishing the overall user experience through further actions required and extra time needed for authentication. Despite these hurdles, a secure solution that is easy to use finally exists.
One of the front-runners in the spoof-proof race is ImageWare, launching Biointellic, an intelligent anti-spoofing technology with all the necessary features to halt presentation attacks. In addition to having proprietary technology that detects liveness better than any other software, this multi-patented service is a server-side solution. Not only are all analysis performed through ImageWare’s servers, but it also works across multiple devices with the same original biometric inputs. The whole authentication process is frictionless, highly scalable, and works with the camera on any standard smartphone.
The industry-leading liveness-detection technology developed by ImageWare is the most advanced solution in the market yet. It combines soon-to-be NIST certified anti-spoofing technology with a zero-friction user experience - the best of both worlds. Facial liveness started as biometrics’ awkward middle child, but now it has the potential to be its superstar.