This work comprises two components: home security and reconnaissance, as well as remote entryway exchange with a greeting framework.  The first step for the user is to install the security and surveillance hardware components of the system.  A growing number of people are considering installing smart home systems as a result of the proliferation of IoT devices.  Security and access control are important components of the smart home system.  This paper details the development of a facial recognition security framework that makes use of Raspberry Pi, which is regularly integrated with smart home systems.  An SDL system that relies on face authentication and intrusion detection is the goal of this project, which intends to use machine learning architectures and the internet of things (IoT).  To secure this smart door, biometric authentication based on face recognition is utilized.  Next, we employ the DeSGCBQNN, which stands for distributed encoder Shannon-Gaussian Correlation Bayesian Q-neural networks, to spot outliers.  We completed the trial examination for several smart entrance face datasets in terms of mean square error, false acceptance rate, false rejection rate, mean average precision, and accuracy.  The proposed approach achieved a 98% success rate, a 66% mean average precision, a 65% false acceptance rate, a 55% false rejection rate, and a 53% mean square error.