Abstract:
The ongoing worldwide apprehension regarding home security emphasizes the enduring significance of personal and familial safety. The widespread increase in home invasions and intrusions can be largely attributed to the absence of standardized security systems in residential areas. In spite of the fact that modern home security systems incorporate video cameras and motion detectors to deter unauthorized entry, the persistent obstacle is accurately distinguishing between authorized occupants and potential threats and preventing false alarms. This study presents a novel resolution that overcomes the limitations of current home security systems through the utilization of Artificial Intelligence (AI) and the Internet of Things (IoT). The proposed system seeks to enhance residential security by integrating object detection and camera technologies, which would enable it to differentiate between authorized residents and unauthorized visitors. One notable characteristic of the proposed system is its ability to engage cameras for accurate image capture and recording, thus augmenting security protocols through the precise detection of motion. By analyzing gestures, images, and movement patterns, the system distinguishes between householders, their families, and visitors, including potential intruders, by employing sophisticated AI models, such as deep learning models. The neural network, more precisely the Convolutional Neural Network (CNN), enhances the system's capacity to detect unauthorized access and reduce the occurrence of false alarms through the examination of camera-captured images. Utilizing image analysis and facial pattern recognition, the study finding has a remarkable 85-100% accuracy of the Convolutional Neural Network (CNN) used and the proposed security system adeptly discriminates between unauthorized individuals and occupants, significantly bolstering the safeguarding of residential establishments. This study, through the formulation of a swift and accurate methodology for discerning authorized from unauthorized individuals and promptly alerting property owners or occupants, seeks to make a noteworthy scholarly contribution to the domain of home security. The proposed solution aspires to surpass current systems in terms of reliability, usability, and efficacy, marking a substantial advancement in residential security technology.
