A Privacy-Preserving Edge Intelligence Framework for Real-Time Multimodal Threat Detection in Smart Urban Surveillance Systems

Jubber Nadaf, Amol K., Vinayak Patil, P. Mathiyalagan, S. K. Lokesh Naik, Indira R.

2025

Abstract

Amidst increasingly complex urban safety challenges, the demand for intelligent, scalable and privacy-aware surveillance system has become urgent. In this paper, we present a new edge computing architecture for the real-time multimodal threat detection in smart cities. By fusing lightweight deep learning models onto the edge device, the solution brings in place video analytics at the edge able to identify unusual behaviors, object defacement, or intrusions with no or little delay. Compared with traditional cloud-based models, data security is guaranteed due to the on-device learning process in the proposed model, and dynamic adaptability to dense and unpredictable urban environments is also provided. The empirical results reveal that the proposed system achieves high threat detection accuracy with a manageable low computational cost, thus indicating its potential in enabling it on diverse smart urban infrastructures.

Paper Citation

in Harvard Style

Nadaf J., K. A., Patil V., Mathiyalagan P., Naik S. and R. I. (2025). A Privacy-Preserving Edge Intelligence Framework for Real-Time Multimodal Threat Detection in Smart Urban Surveillance Systems. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - Volume 1: ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 117-123. DOI: 10.5220/0013858400004919

in Bibtex Style

@conference{icrdicct`2525,
author={Jubber Nadaf and Amol K. and Vinayak Patil and P. Mathiyalagan and S. Naik and Indira R.},
title={A Privacy-Preserving Edge Intelligence Framework for Real-Time Multimodal Threat Detection in Smart Urban Surveillance Systems},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - Volume 1: ICRDICCT`25},
year={2025},
pages={117-123},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013858400004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - Volume 1: ICRDICCT`25
TI - A Privacy-Preserving Edge Intelligence Framework for Real-Time Multimodal Threat Detection in Smart Urban Surveillance Systems
SN - 978-989-758-777-1
AU - Nadaf J.
AU - K. A.
AU - Patil V.
AU - Mathiyalagan P.
AU - Naik S.
AU - R. I.
PY - 2025
SP - 117
EP - 123
DO - 10.5220/0013858400004919
PB - SciTePress