An Automated and Accurate Video Surveillance for Fast Violence Detection Using Machine Learning

K. C. Rajavenkatesswaran, P. Abinaya, T. Kavinkumar, V. Libica, E. Nihethan

2025

Abstract

The main objective of this project is to create a real time fight detection system based on deep learning and computer vision which can enrich the public safety in high-risk areas. But traditional CCTV surveillance has human errors as well as inefficiencies, thus automated surveillance is necessary. It uses YOLO for fast and precise object detection, live streams of video through a Flask backend system, extracts geolocation from live video traffic, and immediately triggers alerts. The system provides real time surveillance and sends alert mail to authorities with incident images at given timestamps and location. It seamlessly integrates into existing security frameworks by reducing manual surveillance efforts. In addition, the system itself was designed to be scalable for deployment in schools or transport hubs or other public places. Features that will be added in future are multi camera integration, sound-based violence detection and predictive analytics working with AI for proactive proactive crime prevention, a safer environment altogether. The public safety in high-risk area is an increasingly important concern as traditional CCTV based surveillance suffers from prosities such as being form humans such as being fatigue, delayed response, and misinterpretation of events. This project presents a real time fight detection system based on deep learning and computer vision, which can help in automation of violence detection and a faster and more easy response to security threats. Based on this, the system is using the YOLO (You Only Look Once) algorithm for real time and highly accurate object detection, which is able to analyze live video streams and identify the violent activities in real time.

Paper Citation

in Harvard Style

Rajavenkatesswaran K., Abinaya P., Kavinkumar T., Libica V. and Nihethan E. (2025). An Automated and Accurate Video Surveillance for Fast Violence Detection Using Machine Learning. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 391-400. DOI: 10.5220/0013898900004919

in Bibtex Style

@conference{icrdicct`2525,
author={K. Rajavenkatesswaran and P. Abinaya and T. Kavinkumar and V. Libica and E. Nihethan},
title={An Automated and Accurate Video Surveillance for Fast Violence Detection Using Machine Learning},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={391-400},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013898900004919},
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 - ICRDICCT`25
TI - An Automated and Accurate Video Surveillance for Fast Violence Detection Using Machine Learning
SN - 978-989-758-777-1
AU - Rajavenkatesswaran K.
AU - Abinaya P.
AU - Kavinkumar T.
AU - Libica V.
AU - Nihethan E.
PY - 2025
SP - 391
EP - 400
DO - 10.5220/0013898900004919
PB - SciTePress