Monitoring the crowd of people by deep learning enabled image analytics

• Main Author: Li, Jiani
• Other Authors: Jiang Xudong
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision; Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/150321
• Institution: Nanyang Technological University

There is a great demand for crowd counting in some practical applications nowadays, such as traffic monitoring, traffic management, sports events and political meetings. In some cases, it is extremely important to obtain information on the number of people. In recent years, many methods and network models for calculating population density have been proposed and made significant progress. However, due to the uneven distribution, high congestion, chaos and occlusion, the effect of the traditional method is not ideal. And the display of the density map is more suitable to meet the demand of real applications. The convolutional neural network can perform well regression and a density map of crowd can be generated by taking the entire image as the input. Based on this method, the functions of accurate crowd statistics and high-quality density map generation are researched and implemented in this project, and a crowd monitoring system based on deep machine learning was developed.