Person Activities Tracking And Detection Using Surveillance System

  The proposed intelligent surveillance framework successfully integrates artificial intelligence, deep learning, and computer vision techniques for real-time person detection, tracking, and recognition using CCTV/IP camera feeds. The framework utilizes YOLOv8 for accurate person detection, DeepSORT for continuous multi-object tracking, InsightFace for facial recognition, and cosine similarity for identity verification, while OpenCV and RTSP streaming enable efficient real-time video processing and surveillance monitoring. The implementation results demonstrate that the system can accurately detect and track multiple individuals simultaneously while maintaining stable identity preservation across consecutive video frames. The facial recognition module effectively identifies target individuals using uploaded reference images and supports frontal-face, side-profile, and partially visible face recognition under varying environmental conditions. The framework significantly reduces manual monitoring effort, improves surveillance automation, and enhances security operations in environments such as colleges, offices, airports, transportation systems, and public places. Although the system performs efficiently in real-time conditions, factors such as extreme occlusion, low-resolution surveillance feeds, and poor illumination may slightly affect recognition accuracy and tracking stability. Future enhancements may include edge AI deployment, cloud-based surveillance integration, low-light optimization, and advanced behavior analysis techniques for improving large-scale intelligent surveillance applications.

[1]   S. C. Tekkök, M. E. Soyunmez, B. Bostancı, and P. O. Ekim, “Face Detection, Tracking and Recognition with Artificial Intelligence,” IEEE, 2021.   [2]   Saket Ranjan, Sarthak Tyagi, Shubhi Gupta, Mandeep Kaur, and Mayank Kumar Goyal, “Image Processing-Based Real-Time Detection and Tracking of Human,” 2023 3rd International Conference on Technological Advancements in Computational Sciences (ICTACS), IEEE, 2023.   [3]   J. Redmon and A. Farhadi, “YOLOv3: An Incremental Improvement,” arXiv preprint arXiv:1804.02767, 2018.   [4]  Glenn Jocher et al., “YOLOv8 by Ultralytics,” Ultralytics, 2023.   [5]  N. Wojke, A. Bewley, and D. Paulus, “Simple Online and Realtime Tracking with a Deep Association Metric,” IEEE International Conference on Image Processing (ICIP), 2017.   [6]   J. Deng, J. Guo, N. Xue, and S. Zafeiriou, “ArcFace: Additive Angular Margin Loss for Deep Face Recognition,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019. [7]   Jiankang Deng et al., “RetinaFace: Single-stage Dense Face Localisation in the Wild,” arXiv preprint arXiv:1905.00641, 2020.   [8]  Guo, Jia, et al., “InsightFace: 2D and 3D Face Analysis Project,” GitHub Repository, 2022.   [9]  G. Bradski, “The OpenCV Library,” Dr. Dobb’s Journal of Software Tools, 2000.   [10]  A. Paszke et al., “PyTorch: An Imperative Style, High-Performance Deep Learning Library,” Advances in Neural Information Processing Systems (NeurIPS), 2019.