How does facial recognition on iPhone X actually work?
By GB Blog Official 2017-11-13 9415 4
How does iPhone X's Face ID work?
Apple's Face ID works in much the same way as fingerprint recognition: it compares two facial images and deducts how similar they are.
The technology behind uses a combo of light projectors and sensors to take pictures of your face — more specifically, facial features. To enable Face ID, Apple uses its TrueDepth camera system: the two technologies work together to create a detailed map of the user's face.
How to set up Face ID on your iPhone X?
As with all Apple devices, setting up the new feature is quite straightforward. All you need to do is follow the on-screen instructions.
You will be asked to move your face in a circle letting the camera take several pictures of your face and create a 3D map. The camera uses infrared light (IR) to light up your face, which allows Face ID to work under different unfavorable lighting conditions both indoors and outdoors: whether it's the lack of light or too much of it.
Here's what happens during image capture:
● The TrueDepth camera (with the help of the proximity sensor and ambient light sensor) determines how much light is needed for correct facial recognition;
● The flood light releases IR (infrared light) to illuminate your face (the light is invisible to the naked eye);
● The dot projector produces enough light to create a 3D map of your face;
● The infrared camera captures images of the dot pattern and infrared light reflected from your face.
That's how all the details of your face are captured and mapped — but how are they later recognized when you want to unlock your phone.
How does the iPhone X recognize your face?
The technology that makes facial recognition possible is based on biometrics. Biometrics are also behind the modern tech's ability to read fingerprints, recognize voices and irises of the eyes. The way all biometric authentication systems work is by comparison. The system compares two patterns and tries to find similarities between them. The patterns may be in the form of the waves of your voice, the structure of your iris or, in case of the iPhone X, the map of your face.
When you want to unlock your phone, the phone will capture the image you are offering and compare it to the template of your face that has been set earlier. The computing then relies on the score from 0 to 1 to make the final verdict. If the score approaches 1, the phone considers it as the same face and unlocks the phone. If it's close to 0, the verification process won't be successful.
Naturally, the verification image and the template image stored in the phone may have been captured under different lighting conditions. The phone takes that into account and there is a different "threshold" set for different scenarios. It is lower when you are just unlocking the phone (for instance, you can pass with a similarity score of 0,7) and higher when you are making a pricey purchase.
Here's what happens during facial recognition on iPhone X:
● The infrared images are transferred from the phone's camera to iPhone X's Neural Engine where a 3D map of the face is created.
● The created 3D map then goes through the processor's algorithms and is compared to the template that was stored in the phone's memory earlier.
● The computer then deducts whether the two images match (the deduction is based on the comparison score described above).
How accurate is facial recognition?
The accuracy of facial recognition depends on three main factors: lighting, pose and facial expression. If all of these conditions are favorable: good illumination, neutral expression, frontal pose — the accuracy of facial recognition can reach up to 99.9%.
An interesting thing to note is that iPhone X's facial recognition has been able to overcome a number of restrictions on other forms of ID. The phone promises to recognize your face even if you wear glasses or a hat or grow a beard.
|You may also want to read:|
|SAR value checklist for iPhone|
|5 best Qi wireless chargers for iPhone X|
|The best alternative phones to iPhone X with similar key features|
2018-10-19By Zim Watson
2020-04-29By Lydia Scott
2017-02-18By Zim Watson
2020-05-05By Joe Horner
2017-12-11By Adeline Belluz
2020-04-27By Cheryl Evans
Prev article:Fix Anet A8 3D printer motor & limited switch issues
Next article:Solved: the Android phone can't add a fingerprint