Use of mobile phones in restricted areas like examination rooms, offices and even institutional premises has become a general issue and it may influence discipline, security and good underlined conduct. Majority of the current options that can be applied to manage this problem are based on manual observation or signal-based classification of the problem. Nevertheless, these solutions cannot work in case of airplane mode used or when no network is connected to the mobile phone. This project proposes a real-time mobile phone detection system based on a computer vision and deep learning as a solution to this issue. Through the system proposed, a yearly camera output is used to identify persons and mobile phones using the YOLOv5 object detection model. Only when a person and a mobile phone are spotted in the same frame, a violation is identified and, therefore, it contributes to minimizing false detainments. An alert is sent to alert a body in case of a violation. The system is also deployed on Raspberry PI 5, which is compact, inexpensive, and can be deployed on edges. There is experimental evidence that the system is capable of operating during real-time and giving good detecting accuracy.
Bhavana et al. (2026) studied this question.