A Multi-Source Surveillance System for Face Detection and Identity Matching Using Live and Pre-Recorded Video Streams

Abstract

This research presents the design and implementation of a multi-source surveillance system for face detection and identity matching using both live CCTV feeds and pre-recorded video streams. The system aims to automate the monitoring process by detecting human faces in video inputs and matching them against a predefined database of known individuals. By integrating computer vision techniques with a user-friendly interface, the system provides real-time and offline analysis capabilities. The proposed system utilizes video processing techniques to extract frames from input sources, followed by face detection and feature encoding. Detected faces are compared with stored facial data to identify individuals and classify them into predefined categories such as authorized, watchlist, or unknown. The system is implemented using Python-based tools, enabling efficient processing and visualization through an interactive dashboard. Experimental results demonstrate that the system can accurately detect and recognize faces under varying conditions, although performance may be affected by lighting, resolution, and camera angles. This approach highlights the potential of intelligent surveillance systems in enhancing security monitoring while also emphasizing the importance of ethical considerations and data privacy. The system can be further improved by incorporating real-time alert mechanisms, larger datasets, and advanced deep learning models to increase accuracy and scalability.