Research on Intelligent Access Control System Based on Interactive Face Liveness Detection and Machine Vision
Wenfeng Shi, Jun Li, Yuanjun Ding, Kuan Jiu Zhou · IOP Conference Series Materials Science and Engineering · 2019
Abstract With the improvement of people’s safety awareness and the renewal of smart home technology, access control security has become a special concern of community, schools and enterprises. How to achieve efficient, convenient, safe and intelligent access control management is the main research direction in the field of smart home right now.Based on multi-task cascade convolutional neural network (MTCNN) and improved face recognition algorithm k-Nearest Neighbor (KNN), this paper proposes an interactive face liveness detection method by eye and mouth state.According to this method, an intelligent access control system based on machine vision is designed.The system solves the problem of identity forgery attacks by calling the camera to track the face in real time and issuing randomized action instructions to the user,after confirming that the object being detected is a living body and the face information matching is successful, the door lock will be opened.The experiment shows that the face recognition rate of the system can reach 98.3%, which has good practical significance.