A Comparative Study for Video Analytics Techniques based on Video Surveillance Cameras

Document Type : Original Article

Authors

1 Software Engineering, Faculty of Engineering, Egyptian Chinese University, Cairo, Egypt

2 Computer Science Department, Faculty of Computers & Artificial Intelligence, Benha University

3 faculty of computer and Artificial intelligence Helwan university

Abstract

Video analytics has recently become a fast-growing trend due to the heavy reliance on video surveillance cameras in many aspects that offer different applications in various domains, such as asset protection, violence detection, traffic monitoring, .etc. Several studies have addressed a variety of video analytic techniques from different perspectives. However, many challenges have been encountered, and more investigations are still required. This paper provides an exhaustive comparative study to investigate the diverse research efforts presented in video analytics based on surveillance cameras, with a detailed discussion and analysis of their main benefits and challenges. Our findings reveal that deep learning methods outperform traditional approaches in crowd anomaly detection, while real-time performance remains a critical challenge for many advanced techniques. Furthermore, this analytical study presents a taxonomy of video analytics applications and techniques and the primary research gaps and limitations that have been concluded, proposing promising directions for future research in video analytics.

Keywords

Main Subjects

Files

Share

How to cite

Statistics