DLIB, YOLO, and Deep Learning-Based Motion and Anomaly Detection Systems for School Security Platform

Abstract

In recent years, the increase in unwanted incidents, especially at the high school level, has raised concerns about the adequacy of school security. This study presents the design and implementation of an artificial intelligence-based system developed to enhance safety in educational institutions. Developed using the Python programming language, the system utilizes OpenCV, YOLO, and the face-recognition libraries to identify students’ faces and monitor movements within the school environment. Each student’s facial data is registered in the system through five photos taken from different angles; subsequently, students’ locations and activities are tracked in real-time via multiple camera streams. The system detects recognized faces and stores them in categorized data files for analysis purposes. Additionally, using a YOLO-based AI model, the system detects prohibited items such as phones, cigarettes, and knives within the school environment, and records these detections along with time and location information in dedicated files. The developed system offers a data-driven approach to addressing incidents at school and contributes to creating a safer and more controllable educational environment.