Real-Time Intrusion Detection and Cloud-based Facial Recognition for Smart Home Security Systems
Josephine Selle Jeyanathan, Cross T. Asha Wise, O. Ansel Tom, C. Yogeswar, G. Sankeerth, Y. Dharani Kumar · 2025
This research presents a comprehensive approach for home security using realtime face recognition, implemented with a Raspberry Pi 5 and Firebase integration. Images of the individuals approaching the door are captured using Rpi camera module in various angles, and through pre-trained CNN model, face detection is performed for authorized access into the house. Known faces automatically unlock the door, while unknown faces trigger an alert to the homeowner’s telegram app with the intruder’s photo and timestamp. The proposed system uses telegram app, integrated with Firebase, provides control on door operations from remote, stores data in cloud, gives real-time notifications, and maintains an accessible log of intruder events. Automated logging ensures that every access attempt is documented for future reference. Experimental results highlight the robustness and reliability of the system, showcasing significantly better accuracy rate (97.2%) using ResNet CNN for detecting the face.