Development of a Smart Home Security System Using Facial Recognition Method Through Convolutional Neural Network

In today's technologically advanced and interconnected world, the significance of home security systems cannot be overstated. Protecting both property and the well-being of occupants has become a paramount concern. Previously, people used traditional methods for their home security systems. These traditional systems relied on the use of external things such as keys, passwords, and ID cards to unlock their homes. By using things like physical keys and ID cards, these items can easily be lost or stolen by irresponsible people which makes the security system of the house vulnerable. To address these identified challenges, this research aims to present a scientific approach by proposing a design for a smart home security system that leverages facial recognition technology. Our approach involves the development of a smart home security system utilizing Convolutional Neural Networks (CNN) for accurate facial recognition. The system incorporates the ESP32-CAM camera and a CNN model, enabling the precise identification of registered faces stored in the system's database. Through experimentation, the proposed prototype based on the FaceNet CNN achieved exceptional accuracy of 99.26%. This demonstrates the effectiveness of the developed smart home security system. The ESP32 camera successfully detects registered faces, recording the identities of individuals along with corresponding timestamps. These positive outcomes highlight the system's potential in enhancing the security of smart homes. Overall, this research not only addresses the pressing issue of home security but also contributes to the advancement of scientific knowledge in the field of smart home technology. The proposed design and implementation of the smart home security system paves the way for improved security measures in the context of modern residential environments.