Facial Recognition

iPhone X

With the new iPhone X [pronounced: ‘Ten’] Apple has brought us into a new age of digital security. Up to now we have seen movies where people enter secure areas by having the iris of their eyes scanned or hand prints examined, and of course the fingerprint has been there for a good many years. But now we are into the Face ID era.

The latest Microsoft Surface Pro 4 tablet came with a secure login programme where the authorized users can introduce themselves to their tablet, and thereafter it will look for them, logging them in when it recognizes their face, or closing down when they do something like take a lunch break without logging off. There is a third camera hidden in the bezel and because it’s infrared, it even works in dim and nearly dark environments. It is still possible to log in using a password just in case you aren’t keen on the facial recognition trick.

Now Apple has gone a step further, three-dimensional measurements of your face and this is how it works.

The iPhone uses infra lights and AI [artificial intelligence] to detect your face’s unique shape and match it to the stored version in its memory.

Bear in mind that iPhone X is only just now being delivered to buyers in this country and Apple is not about to tell me or anybody else how their proprietary system works, however this is the educated guess of those most in the know.


PX Face ID

1. First up the phone detects your gaze –  yes, it is ‘attention aware’, it waits for the moment you are looking directly at the screen.

2. The flood illuminator sends out a burst of invisible infrared light.

3. The dot projector sends a vertical cavity surface-emitting laser through a lens, dot spreadprojecting 30.000 points of infrared light onto your face. It is easy to imagine that this makes big computational demands and if activated as frequently as the Flood illuminator could rapidly drain the battery.

4. Next a random sample is sent to a neural network [advanced number-crunching software designed to mimic the human brain, trained with more than a billion facial images] on Apple’s custom A11 Bionic chip.

5. The neural network compares the new face data with the established facial data [it is a mathematical representation of your face snapped across a variety of poses]. And there is a second neural network to spot and resist “spoofing” using photos or masks.

6. After all this and if everything matches your phone is unlocked. If not the apps using Apple’s security system are enabled.

Apple says that Face ID will even learn how your facial hair grows, if you happen to grow facial hair, and it will adapt accordingly. It will also work in complete darkness, and if you are wearing a hat, glasses or a scarf, just as long the infrared beam can pass through it to reach your face.

So what’s the problem? Just that to unlock your phone you have to give it your attention, you have to look at it, really look at it. You can’t surreptiously glance at your phone whilst talking to somebody, you have to give it your attention.

What are the chances than someone else can unlock your phone using Face ID? They say 1 in 1,000.000 compared to the probability that a person can unlock an iPhone with Touch ID being 1 in 50,000.

It is likely that facial recognition will grow in use in the future given the enhanced security it affords, and even if you have an evil twin there is still enough facial differences from natural ageing and weight differences, but just in case you are concerned use a pass code to authenticate.

Now it is only the cost you might worry about.

Happy communicating!

1 thought on “Facial Recognition

  1. Isabel,
    I cannot change your registered email address on WordPress (our website host). You will need to do that yourself by deleting the @paradise.net.nz address and re-adding your new @gmail.com address. Wish I could help but the WordPress System won’t let me touch registered email addresses (other than deleting them!).

Comments are closed.