A System for Detecting Dangerous and Suspicious Human Behavior for Educational Institutions

This study aims to minimize the security vulnerabilities that may occur in educational institutions by using deep learning solutions. The designed system detects cutting and piercing tools to distinguish between dangerous, suspicious, and normal human behaviors in indoor and outdoor spaces of educational institutions. Thanks to the detected behaviors, it is ensured that the students in the institution reach a safer education model. In this project, where students are provided with safety education, Deep Learning Neural Network Model is trained and human behaviors are categorized with the images obtained with the help of the camera. Pistols and knives can be detected with over 90% of successful predictions.