DEVELOPMENT OF A FACE DETECTION SYSTEM FOR SUSPECT ALERT VIA CLOUD COMPUTING USING INTERNET OF THINGS TECHNOLOGY

Abstract

This research presents the Development of a Face Detection System for Suspect Alert via Cloud Computing Using Internet of Things Technology  (IoT). The objectives of this study were: (1) to develop a suspicious person notification system, (2) to develop a cloud-based storage system for recording suspicious person information, and (3) to enable the system to send real-time notification messages to homeowners via Line Notify whenever a suspicious person is detected.

          The proposed system was implemented using a Raspberry Pi 3 Model B+ as the central processing and control unit, together with a Full HD Auto Focus Webcam A881 for image acquisition. During system operation, facial images of individuals entering the monitoring area are captured and transmitted to the Raspberry Pi for image processing and facial recognition. The system was developed in Python, using the Thonny Python IDE as the development environment.

          The system performance was evaluated through questionnaires and performance assessments involving three experts and fifty general users residing in Sai Mai District, Bangkok, selected through simple random sampling. Experimental results demonstrated that the developed facial recognition model achieved high predictive performance, with an overall accuracy of 98.74%, a precision of 98.61%, and a recall of 98.52%. Furthermore, the experts' evaluation indicated a mean score of 4.61 (S.D. = 0.58), representing a very high level of system performance. Similarly, the user satisfaction assessment yielded a mean score of 4.57 (S.D. = 0.60), also corresponding to a very high level of satisfaction. The findings indicate that the proposed system provides highly accurate facial recognition, effectively supports real-time monitoring and notification, and successfully fulfills all research objectives.