The system measures the facial hole and generates a facial pattern, which is a unique numerical code for the face. After storing the facial pattern, the system will compare it with thousands or millions of facial pattern data stored in the database. Each facial pattern is stored as an 84 byte file. The system can match facial pattern data in memory at a speed of 60 million faces per minute, and for facial pattern data in the hard drive, it can match 15 million faces per minute. Operators can view two photos declared as matches to determine if the computer's operation is accurate.

After the facial recognition system is connected to the video surveillance system, the recognition software will search for facial information in the camera's field of view. If there is a face in the field of vision, it will detect it within a fraction of a second. It uses multi-scale algorithms to search for facial images at low resolution. An algorithm is a program that provides a set of instructions to complete a specific task. The system only switches to high-resolution search after detecting a shape similar to a head.