The TrueDepth Camera System uses Infrared light when capturing images of your face. This way, the system can instantly recognize you under any light condition (day or night) and at any place (indoor or outdoor).  Of course, the IR light compensates any background effects or lighting. The following are steps of image capture as a way on how Face ID works:
  • The ambient light sensor and proximity sensor in the TrueDepth Camera System calculate the illumination required for face recognition
  • The illuminator emits Infrared light to illuminate your face
  • The projectors create more than 30.000 dots used to map your facial landscape
  • The camera captures images, which are combinations of dot patterns and IR light.
Face Recognition
Human face is a measureable biological characteristic. Just as the technology used in fingerprints or voice recognition, face recognition system works by comparing two complex patterns and determining how authentic they are. In terms of Face ID, the complex patterns refer to the landscape of your facial features.
Then, the processor in the smartphone will capture the reference patterns and store them as a template. When you want to unlock the phone by using Face ID, the system will compare your face and the stored template, and then calculate the similarity. If it finds that it is the same person as you, the smartphone is unlocked automatically.
Of course, you cannot capture images of your face at identical conditions. For this, the phone uses a threshold to determine whether the images are significantly similar or not. The threshold ranges from 0 to 1. Zero score means that that the face on the screen is not similar to the stored template. On the other hand, 1 means perfect authenticity. The scores of 7 or 8 are mostly identical. They are reliable enough to unlock the phone.  However, for payment authentication, the threshold may be higher, 0.9 for instance.
Authentication Accuracy
Accuracy of the Face recognition depends upon three variables, namely, pose, facial expression, and illumination. When the pose is identical, the illumination is uniform, and the facial expression is identical, the authenticity can be as high as 99,99%. However, when your face is obscured because of inadequate illumination or because of something that covers up your face, the authentication accuracy will be lower.
Face recognition technology, which has been used for various purposes like Visa, driver’s license, or passport, has posed some problems. This is because the government applications are strict. As a result, the photo must be taken at a standard pose, and you cannot laugh or even smile, the background is set in certain color, and you cannot use any facial accessories like glasses.
The Face ID used in iPhone X seems to overcome those problems.  Apple’s representative said that the Face ID is built upon the cutting-edge technology, which enables the smartphone to recognize your face despite changes in expression, pose, or illumination. It even works when you wear facial accessories like hats and glasses or when you have different haircuts.