Apple's FaceID has long been able to carry out deep learning and simulation according to the user's appearance changes. Even if you wear glasses, make up and grow a beard, you can gradually simulate the image through autonomous learning, which will be faster and more accurate in the next recognition.
For a big appearance change like wearing a mask, the iPhone may not recognize you for the first time and need to enter a password.
If FaceID unlocking fails for dozens or even hundreds of times in a row, after the password is successfully unlocked, iPhone will slowly learn your appearance to record the user's new appearance, so as to re-remember your appearance and recognize that you are wearing a mask.
Extended data
FaceID needs huge sensing equipment and supporting equipment carried by iPhone in the face recognition process, such as infrared camera, distance sensor, ambient light sensor and so on. According to a patent submitted by Apple, the general process of FaceID in the identification stage is as follows:
In the process of face recognition, iPhone will call all sensors to get face information at the same time, and 3D sensors are responsible for identifying information such as eye socket contour, chin, nose, skin lines and even spots on the skin.
Infrared sensor is responsible for identification in dark environment. Through sensors and middleware, devices can perceive the surrounding three-dimensional environment and realize face recognition, which is also one of the key technologies involved in FaceID.