DOI: 10.35377/saucis...1805817 ISSN: 2636-8129

Internet of Things-Based Smart Security System with Face and Object Detection Using Machine Learning

Benjamin Ovioisa, Hüseyin Güney
Show PDF Cite
The rapid advancement of the Internet of Things (IoT) has enabled significant progress across various sectors such as finance, healthcare, smart homes, and smart cities. One key application is the development of smart security systems, which are gaining traction due to their efficiency, reduced need for human input, and enhanced threat detection. However, many existing systems face challenges such as limited coverage, false alarms, weak authentication, privacy concerns, slow response times, and dependency on external resources. To overcome these issues, this paper introduces an IoT-based smart security system that uses machine learning for facial and object recognition. The system employs a Convolutional Neural Network (CNN) to detect faces and a Single Shot Detector (SSD) for identifying suspicious objects. When an unknown individual is identified, a push notification alerts the administrator for further action. The system demonstrated high performance, with CNN achieving 94\% accuracy and an F1-score of 95.24\%, while SSD achieved 90\% accuracy and an F1-score of 94.7\%. This intelligent security solution has potential applications beyond the scope of this study and could be effectively implemented across multiple industries, enhancing safety through advanced technology and AI-driven methods.

More from our Archive