Development of a Child-Adult Binding Access Control System Using Face Recognition

Aiming to address the need for child supervision in closed public spaces, this paper proposes an intelligent access control system based on face recognition. The system adopts lightweight MTCNN and MobileFaceNet algorithms to dynamically bind the identities of children and adults by generating unique ID codes (timestamp + random six-digit number). During entry, an adult generates an ID code via face detection and associates it with the child's information. During exit, a child can only pass if their ID code matches the associated adult's ID. Experimental results demonstrate that the system achieves a recognition accuracy of 99.93%, a single binding time of 2.8 seconds, and a database response time of less than 10 ms under high concurrency (1000 QPS). This system enables accurate guardianship without pre-registration and effectively reduces the risk of children getting lost or being taken by unauthorized individuals.