Video based human crowd analysis using machine learning: a survey. Issue 2 (4th March 2022)

• Record Type: Journal Article
• Authors: Chaudhary, Deevesh; Kumar, Sunil; Dhaka, Vijaypal Singh
• Is Part Of: Computer methods in biomechanics and biomedical engineering. Volume 10:Issue 2(2022)
• Page Start: 113
• Page End: 131
• Publication Date: 2022-03-04
• Subjects: Crowd behaviour -- crowd analysis -- crowd features -- crowd tracking -- CNN; Imaging systems in biology -- Periodicals; Imaging systems in medicine -- Periodicals; Biomechanics -- Data processing -- Periodicals; Biomedical engineering -- Periodicals; 616.0757
• DOI: 10.1080/21681163.2021.1986859
• Languages: English
• ISSNs: 2168-1163

Abstract

ABSTRACT: World population has increased manifolds in the last ten years. With the increase in population at this alarming rate, studying and understanding crowd patterns and their collective behaviour is very important. Researchers from various domains like artificial intelligence, machine learning, social science have shown their interest in understanding crowd phenomena from the social, psychological, and technical points of view. Computer vision techniques play a vital role in developing methods that help in understanding and analysing crowd behaviour automatically. In this article, we have surveyed many models related to crowd analysis developed and employed in computer vision. We aim to provide a comprehensive overview of the research from different aspects of crowd analysis like crowd count, human detection, anomaly detection, human behaviour, the importance of crowd analysis, and recent developments in this field. Major contributions have been included, along with their strengths and limitations.