Real-time face recognition and prevent unauthorized entry on campus using pre-trained models and opencv

Abstract

Biometrics, involving the analysis of human behavior and characteristics, has become instrumental in enhancing security measures. This paper explores the various approaches employed in face recognition, a pivotal biometric technique. The survey conducted here evaluates different algorithms and processes, positioning face recognition as a novel biometric division with promising applications in healthcare security. The focus shifts to the utilization of face detection methods with computer assistance. Shifting gears, the paper delves into a general theory overview, juxtaposing behavioral and neural phenomena in face recognition against object recognition. Original experiments are presented to test key assumptions, highlighting the need to discern and apply Convolutional Neural Network (CNN). The overarching goal is to revolutionize traditional security systems in educational institutions and restricted areas. This proposed system aims to streamline verification processes, significantly reducing the time required for traditional campus security measures. The system excels at identifying individuals not registered in the database or unauthorized personnel, thereby fortifying office security against potential thefts. As the research unfolds, it becomes evident that this innovative approach to security stands as a beacon of efficiency, offering a swift and reliable means of verification with substantial implications for theft prevention and enhanced overall security. Our achieved accuracy is 96.5% in face recognition.