Design and implementation of face recognition-based door access control system
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
With recent dramatic development in the field of artificial intelligence (AI), smart access control has become crucial part of our modern everyday lives. This paper presents a door security system designed to prevent trespassing in a highly secure areas like home environment. The implemented system is cheaper and more reliable for intruder detection and door security. The door access is solely based on face recognition to ensure that only authorized individuals go through the door. The principle on which face recognition works is that the sysadmin defines an individual or a group of individuals who are allowed to enter the area, a room or a building, through a door, a gate, or any physical barrier. Thus, the access is limited to these individuals. Faces of authorized individuals are captured, stored, and trained by the system, and a real-time face is captured and matched against the ones of authorized individuals to identify the individual gaining access through the door after which the access is granted if that individual is authorized and denied otherwise. Haar-based cascade classifier was used for face detection while Local Binary Pattern Histograms (LBPH) algorithm was used for face recognition. 150 faces of individual were captured and trained after which real-time faces were tested on the system to determine the accuracy. The implemented recognition system achieved a recognition or accuracy rate of 88.6%. Notification has been achieved, using Simple Mail Transfer Protocol (SMTP), by sending emails to the sysadmin in case of any successful or unsuccessful attempt to gain access through the door. Python’s Open Source Computer Vision (OpenCV) library was used for face recognition. The implemented system is cheap, user-friendly, and reliable.