Abstract

Muhammad Khurram Latif

Student ID: 11670746

Project Title: Crowd Surveillance and Abnormal Behavior Detection System

Security has become a major challenge all over the globe these days. So, crowd surveillance with video cameras is a key factor to maintain security especially in public and crowded areas. But it is quite hard to track the human behavior because actions of human being are usually unpredictable and random. The project research will find a system which captures the data from CCTV cameras to observe and differentiate between the abnormal/anomalous and normal behavior of individuals. The aim and purpose of project is to produce such an intelligent surveillance system that can use to improve the security of public places e.g. educational institutes, parks, grounds, shopping centers, hospitals etc. The system will work in unsupervised approach and be able to detect and predict the abnormal event automatically to control and avoid the big disasters. The system will work on both pre-recorded and live footages. It will provide reliable results after carefully monitoring unusual activities.

Resources:

Nayan, N., Sahu, S., & Kumar, S. (2019). Detecting anomalous crowd behavior using correlation analysis of optical flow. Signal, Image And Video Processing, 13(6), 1233-1241. https://doi.org/10.1007/s11760-019-01474-9

Recent Advancement in Anomaly Detection in Surveillance Videos. (2019), 8(2), 964-970. https://doi.org/10.35940/ijrte.b1759.078219

Zhang, X., Yang, S., Tang, Y., & Zhang, W. (2015). A thermodynamics-inspired feature for anomaly detection on crowd motions in surveillance videos. Multimedia Tools And Applications, 75(14), 8799-8826. https://doi.org/10.1007/s11042-015-3101-8