A Systematic Survey of Abnormal Event Detection Models from Video with Application, Limitations and Challenges

An increase in data accompanies daily advancements in technology. The analysis of data in video format is also a challenging task. Research on methods for event detection in video requires attention from scientists because of its use in many applications such as video surveillance systems, advanced human-computer interaction interfaces, robotic surgery, automatic driving, etc. Event detection and recognition are performed based on low-level features, high-level features, resolution, color, and size. The difficult task in video surveillance is to continue monitoring to find out about some abnormal events, which is very time-consuming and boring work. In this paper, author reviewed previous year's papers for video surveillance in different areas like hospitals, universities, indoor environments, and live sports recording. Author list out some abnormal events, some models and techniques with advantages and disadvantages, and benchmark datasets from different papers. In addition, they draw attention to a few challenges for modern analysts.