The iPhone X A11 Bionic chip has machine learning support, allowing Face ID to learn and remember changes. If you grow beard, cut hair or wear glasses, Face ID will remember the changes. It will have to look at how it actually works but if it is as Apple says it will be a major advance in the use of face recognition technology.
Infrared light allows Face ID to work even in the dark. The camera itself provides light source in the form of infrared rays, so no other external light source is needed. It also has a Flood Illuminator (a form of infrared flash) to balance the light on the face. Because the human eye is not visible, there will be no blinding flash light. Overall, even in the dark, iPhone X can still identify your face.
According to Apple, Face ID avoids phishing attacks that make other face-unlocking systems unsafe. That's because it uses an infrared camera that is sensitive to depth. An attacker cannot trick a camera with a photo. Even 3-dimensional masks cannot.
Although the method to avoid scams has not been announced yet, infrared light can penetrate the skin, allowing TrueDepth Camera to scan under the skin, making masks or images difficult to imitate. It is also possible that many pixels make it more difficult to create masks to trick.
Mr. Craig Federighi introduced Face ID at the iPhone X launch event
Like the Touch ID fingerprint, the face math model will be in the iPhone X's Secure Enclave. They are encrypted and other applications or OS cannot access it. The detected face is only on the device and is not stored on Apple's server or in sync with iCloud. So far, Secure Enclave has not been hacked.
According to Apple, Face ID's error rate is 1 in 1,000,000. It means that Face ID will be unlocked wrong 1 in 1 billion times. Touch ID has an error rate of 1 out of 50,000, making Face ID safer. Like Touch ID, Face ID can authenticate Apple Pay. Does this mean that Face ID will replace Touch ID? If it can be as easy and reliable as Apple says it is possible.
See more:
The future of the iPhone depends on how you think Face ID