AN IOT-BASED FACE RECOGNITION SMART DOOR ACCESS CONTROL SYSTEM WITH REAL-TIME SECURITY ANALYSIS AND PERFORMANCE OPTIMIZATION

Abstract

Smart door access control systems are becoming more significant in modern homes and enterprises as the need for better security, automation, and remote monitoring increases. Older locks have issues with lost keys, keys that have been copied, and unlawful entry. This study provides an IoT-based facial recognition smart door access control system that mixes computer vision, integrated hardware, and cloud technologies for safe and automatic authentication. The suggested system is composed of a vision pipeline that includes face detection, face alignment, feature encoding, face matching, and decision making to identify authorized users in real time. Upon successful authentication, an Arduino-controlled servo motor opens the door, a welcome message is shown on an I2C LCD, and the event is logged in a Firebase cloud database. To enhance security, a liveness detection technique based on eye-blink verification is integrated to limit the danger of spoofing attacks utilizing static facial photos. We performed an experimental assessment, and the results indicate that the system is capable of dependable real-time authentication with minimal processing latency and excellent cloud-based monitoring. The suggested system is a low-cost, safe, and scalable strategy for smart access control applications in home and corporate contexts.