Facial authentication based smart door lock system and anomaly detection using machine learning architectures integrated with IoT

Home security and reconnaissance, as well as far off entryway exchanging with a hello framework are two parts of this work. Process of installing system's hardware components for security and surveillance begins the user's journey. With advent of Internet of Things (IOT), there is an increase in interest in smart home systems in recent years. One of the significant parts of the brilliant home framework is the security and access control. In this paper, a facial acknowledgment security framework was planned utilizing Raspberry Pi which are consistently coordinated to savvy home framework. Using machine learning architectures and IoT, this study aims to develop a smart door lock (SDL) system based on facial authentication and intrusion detection. Biometric authentication that is based on facial recognition is used to lock this smart door. The distributed encoder Shannon Gaussian Correntropy Bayesian Q-neural networks (DeSGCBQNN) are then used to detect anomalies. The trial examination is completed for different savvy entryway facial dataset as far as accuracy, mean average precision, False Acceptance Rate, False Rejection Rate and mean square error.Proposed technique attained accuracy of 98%, mean average precision of 66%, False Acceptance Rate of 65%, False Rejection Rate 55%and mean square error of 53%.