Deep Learning in Smart Video Surveillance for Crowd Management: A Systematic Literature Review

A crowd is defined as a gathering of people in the same premises. When the number of people exceeds normal conditions, overcrowding becomes a concern in safety and health-related matters due to the risks that a large crowd can impose on the individuals present in the area. Crowd analysis is a growing trend in computer vision related to the concerns in crowd monitoring. To alleviate risks related to crowds, intelligent techniques applied to surveillance are used to analyze a crowd and to monitor its density and the behavior of people captured in footage. Through a systematic literature review of various papers published in the last five years related to crowd analysis, the numerous deep learning algorithms applied in past researchers are presented and are assessed to come up with a solution that will further aid in crowd management