Design of an Adaptive Crowd Management Detector System for Massive Events

Authors

  • Manar Salamah Ali, Maha Saud Alarifi, Raneem Yasser Gharbi

Abstract

The paper focuses on complex event recognition techniques for crowd management in massive events, like Hajj. It proposes a computer vision and machine learning system which detects abnormal crowd behavior from online Hajj footage. The system applies an adaptive co-training technique and can adapt to scene changes in Hajj videos. The performance and efficiency of the proposed system is evaluated based on publicly available Hajj video content and it is shown that it provides superior detection accuracy and precision.

Keywords: Hajj crowd management, complex event recognition

Published

2020-07-30

Issue

Section

Articles