Face Authentication System With Liveness Detection Security

As face authentication systems continue to improve, their use is becoming more and more widespread, from catching your next flight to the smartphone in your pocket. According to a report by the National Institute for Standards and Technology (NIST), the best face authentication algorithms now have an error rate of less than 0.2 percent, far better than human performance. However, face authentication technology on its own isn’t foolproof. In order to thwart malicious actors and bolster security, many organizations have taken to using face authentication with liveness detection. But what is liveness detection exactly, and how does liveness detection for facial authentication work?

What is Liveness Detection?

“Liveness detection” refers to an AI system’s ability to distinguish between a live person and a false representation of that person. The task of liveness detection can be used for any biometric system, including face authentication, fingerprint scanning, iris scanning, speech recognition, etc. The possible forms of liveness detection for facial authentication include:

• Analyzing eye or mouth movements.
• Prompting the user to perform a certain action.
• Detecting 3D depth information.

The term “liveness” was coined by computer science professor Dorothy E. Denning in her 2001 article “Why I Love Biometrics,” which is the source of the well-known quote “A good biometrics system should not depend on secrecy.” Denning argued that in order to be considered truly secure, face authentication software and other biometrics systems can’t just depend on hiding their inner operations—instead, they need to be truly robust to external attacks.

How Does Liveness Detection Improve Security?

There are multiple possible ways that an attacker might try to counteract a face authentication system:

• Presenting a photograph or video of the impersonated individual.
• Presenting the individual to be impersonated when they are unaware of it (e.g. while sleeping or looking away from the camera).
• Presenting a false representation of the impersonated individual (e.g. a mask or realistic makeup).

These methods are collectively known as a “presentation attack.” Liveness detection is the task of identifying these attacks and differentiating them from real users’ attempts to access the system, allowing legitimate users to continue to enjoy a fluid, uninterrupted experience.

How Does Liveness Detection Work?

There are two types of liveness detection methods, “active” and “passive”:

• Active liveness detection involves methods that require the user’s active participation. These techniques may include asking the user to follow motion on the screen with their eyes, or move the camera or their body in a particular way. While effective, active liveness detection requires the user to be aware of what the system is asking for, which a skilled attacker may learn to exploit.
• Passive liveness detection involves methods that work without the user’s active participation. These techniques may include examining the face (e.g. its skin, texture, and borders) for signs that it is a false representation (e.g. a mask or cut-out image).

Active liveness detection methods should generally be supplemented by passive liveness detection in order to provide a high degree of confidence in the security of a face authentication system. Some of the forms of passive liveness detection include:

• Texture analysis to detect whether a user’s face is real or a spoof.
• Variable focus methods that compare pixels from images taken with different levels of focus.
• Optical flow algorithms that examine how objects move between two frames of a video.

How Businesses Use Face Authentication with Liveness Detection

Face authentication with liveness detection is useful in a wide range of contexts and industries. Below are just a few examples:

• Healthcare AI: Doctors, nurses, and other healthcare providers can use face authentication systems for authentication (e.g. when viewing confidential patient information, or accessing a restricted area). Patients can also use face authentication to verify themselves, reducing medical errors due to mistaken identity, or to detect emotions and pain visible from their expression.
• Security AI: Face authentication with liveness detection is tremendously valuable in the field of safety and security, making it exponentially more difficult for attackers to spoof another person’s identity. Enforcing liveness detection adds another layer of protection for off-limits areas and locations, from construction sites to office buildings.
• Retail AI: Retail stores can use face authentication systems for the benefit of both employees and customers. Employees can instantly check in and out for the day by scanning their face, while customers can enjoy touchless payments. Face authentication can also be used to identify repeat customers or as a deterrent to shoplifting.

Conclusion

Chooch’s facial authentication with liveness detection solution makes it easy for organizations of all sizes and industries to build rock-solid face authentication into their workflows. With extremely high accuracy and lightning-fast response times, Chooch’s facial authentication AI is a robust, mature solution for your needs and objectives. Get in touch with our team today for a chat about how we can help.