Authors : Pooja Deshmukh; Dr. Smita Ponde

Volume/Issue : Volume 11 - 2026, Issue 5 - May

Google Scholar : https://tinyurl.com/36yczkks

Scribd : https://tinyurl.com/mr28b7a9

DOI : https://doi.org/10.38124/ijisrt/26May402

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Abstract : The proliferation of Internet-connected cameras and the increasing need for intelligent, automated security monitoring have created demand for systems that go beyond simple video recording. This paper presents CCTV Guardian, a web-based real-time surveillance platform that integrates deep-learning face recognition, liveness-based antispoofing via Eye Aspect Ratio (EAR) computed from MediaPipe Face Mesh 3-D landmarks, suspicious activity detection using skeletal pose analysis and Gunnar–Farnebäck optical-flow motion estimation, automated email alert delivery with snapshot attachments, and IP/RTSP camera connectivity—all accessible through a Flask browser-based dashboard. The system stores authorised face embeddings computed with the dlib 128-D ResNet model and performs per-frame recognition at runtime. Spoof attacks are rejected by verifying spontaneous eye-blink patterns within a 6-second temporal window. Unusual body posture (raised hands, crouching) and abnormal motion velocity are flagged as activity alerts. An SMTP email module with a 60-second cooldown dispatches alerts with JPEG snapshots to designated recipients. Tested with 22 enrolled persons across varied lighting conditions and distances, the system achieved a face-recognition accuracy of 96.4%, with anti-spoofing correctly rejecting all 50 static-photograph attacks in evaluation trials. The system maintains 12–15 fps on commodity hardware without a dedicated GPU and is designed for small-to-medium enterprise and residential security applications.

Keywords : Face Recognition; Anti-Spoofing; Liveness Detection; CCTV; Flask; MediaPipe; EAR; Suspicious Activity Detection; RTSP; Email Alert; Optical Flow; Surveillance System.

References :

1. M. Turk and A. Pentland, "Eigenfaces for recognition," J. Cognitive Neuroscience, vol. 3, no. 1, pp. 71–86, 1991.
2. P. Viola and M. Jones, "Rapid object detection using a boosted cascade of simple features," Proc. IEEE CVPR, 2001, pp. 511–518.
3. Y. Taigman, M. Yang, M. Ranzato, and L. Wolf, "DeepFace: Closing the gap to human-level performance in face verification," Proc. IEEE CVPR, 2014, pp. 1701–1708.
4. F. Schroff, D. Kalenichenko, and J. Philbin, "FaceNet: A unified embedding for face recognition and clustering," Proc. IEEE CVPR, 2015, pp. 815–823.
5. D. King, "Dlib-ml: A machine learning toolkit," J. Machine Learning Research, vol. 10, pp. 1755–1758, 2009.
6. J. Li, Y. Wang, T. Tan, and A. K. Jain, "Live face detection based on the analysis of Fourier spectra," Proc. SPIE Biometric Technology, 2004, pp. 296–303.
7. G. Pan, L. Sun, Z. Wu, and S. Lao, "Eyeblink-based anti-spoofing in face recognition from a generic webcamera," Proc. IEEE ICCV, 2007, pp. 1–8.
8. T. Soukupová and J. Čech, "Real-time eye blink detection using facial landmarks," Computer Vision Winter Workshop (CVWW), 2016.
9. C. Lugaresi et al., "MediaPipe: A framework for building perception pipelines," arXiv preprint arXiv:1906.08172, 2019.
10. G. Farnebäck, "Two-frame motion estimation based on polynomial expansion," Proc. SCIA, 2003, pp. 363–370.
11. A. Ronacher, "Flask: A lightweight WSGI web application framework," 2010. [Online]. Available: https://flask.palletsprojects.com
12. M.H. Wani and A.R. Faridi, "Deep learning-based video surveillance system for suspicious activity detection," J. Intelligent & Fuzzy Systems, vol. 47, pp. 71–82, 2024. DOI: 10.3233/JIFS-234365.
13. A. Deshpande and R. Kulkarni, "Real-time face recognition based security systems: A survey," IJEAT, vol. 9, no. 3, pp. 2248–2254, 2020.
14. S. Ghosh and P. Bhattacharya, "Deep learning-based surveillance systems: A review," IEEE Access, vol. 9, pp. 103661–103683, 2021.