November 14, 2017

Apple Says, Face ID Can Learn From It’s Mistakes

iPhone X is the first iPhone from Apple with Face ID technology, a facial recognition system which uses the TrueDepth camera system on the top of the display to unlock the device that has replaced the plain old Touch ID fingerprint scanner.


How Face ID Works?

The TrueDepth camera on the iPhone X is made up of several different elements, all of which combine to make your 3D facial image. An infrared camera, dot Projector, flood illuminator, Apple’s own specialized hardware, proximity and ambient light sensors,  all of which map the face with 30,000 invisible dots flashed on the visage. That information feeds the iPhone X’s neural network, which creates a mathematical model of the user’s face.

The infrared camera then sends the data directly to the ‘Secure Enclave’ within the iPhone X’s A11 Bionic processor chip. Here it is checked against the pre-scanned image to ensure that the correct face is being seen. If a match is found, the phone is unlocked, all within a blink of an eye.




How Face ID learns from its Mistakes?

Although Apple says, Face ID is better than Touch ID with one in a million error rate, there is a still tiniest fraction of a possibility that Face ID might fail. Even at those times, the Face ID will learn from its mistakes using AI to recognize the faces better next time when the user tries to unlock the device, Apple says. All you need to do is unlock the iPhone X using the password to train the Face ID.

According to the White Paper published by Apple, when a user enters the password manually when the Face ID fails, it takes another snapshot of your Face using the True Depth Camera and updates the system with the new image. With the powerful AI (Artificial Intelligence) and neural network-based system developed by Apple, Face ID will regularly adapt to your facial changes even if you grow a beard to reduce the error rates.


“To improve unlock performance and keep pace with the natural changes of your face and look, Face ID augments its stored mathematical representation over time. Upon successful unlock, Face ID may use the newly calculated mathematical representation—if its quality is sufficient—for a finite number of additional unlocks before that data is discarded.

Conversely, if Face ID fails to recognize you, but the match quality is higher than a certain threshold and you immediately follow the failure by entering your passcode, Face ID takes another capture and augments its enrolled Face ID data with the newly calculated mathematical representation. This new Face ID data is discarded after a finite number of unlocks and if you stop watching it. These augmentation processes allow Face ID to keep up with dramatic changes in your facial hair or makeup use while minimizing false acceptance.”

About the author 


{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}