Real-Time Video Surveillance Face Detection System

Abstract

Abstract——The Real-Time Video Surveillance Face Detection System offers a modernized solution for automated surveillance by leveraging Python, Django, and OpenCV to achieve reliable facial recognition in security-sensitive environments. Utilizing cascaded classifiers, as proposed by Jones and Viola [1], this system rapidly identifies faces in real-time, while managing a dynamic profile database. Additionally, Saraswat and Kushwaha's work on CCTV-based face detection [2] informs our approach to handling challenges like low lighting and crowd density. Our methodology combines efficient video processing, profile management, and deep learning-based facial recognition, achieving a 90% accuracy rate under controlled conditions. Our implementation enables security personnel to add, edit, or delete profiles, providing flexibility as security requirements evolve. Tests confirm the system’s adaptability across varied lighting and angles, making it effective for high-stakes surveillance environments. Future improvements will explore multi-camera support, automated alerting, and optimized recognition in low-light scenarios. With these upgrades, this system presents a scalable and reliable tool for modern security infrastructure, advancing the field of real-time video surveillance. Keywords—Face Recognition, Surveillance, Real-Time Detection, Video Processing, OpenCV, Security Systems