Student Behavior Monitoring System in Classroom Environment Using Yolov8 Method

Abstract

Student ethics in an academic environment is an important element in creating an orderly and professional learning environment. One form of ethical violation that is still often found in the lecture environment is eating and drinking activities in the classroom. This study's objective is to develop an automatic detection system for unethical student behavior in the classroom, especially eating and drinking activities, utilizing one of the newest Real-time deep learning approaches object recognition on a Raspberry Pi device, The algorithm known as You Only Look Once version 8 (YOLOv8). A special dataset was developed through a manual annotation process in the form of images and videos showing students with various activities in the classroom. This system is expected to be an additional solution in monitoring student ethics automatically, efficiently, and in real-time in a modern learning environment. The test findings demonstrate that the model can identify eating and drinking activities with a respectable degree of precision indicating that the system is able to detect target activities with an accuracy level of up 95% with fairly stable performance in good lighting conditions and certain viewing angles