Detecting abnormal crowd behaviors based on the div-curl characteristics of flow fields
作者：
Highlights：
• Explore the crowd state change detection problem for the dense crowd scenarios of wide-field of view in the public place.
• Use a less computation required thermal diffusion processing and perspective transform processing to capture the macroscopic crowd motion information.
• Model motion information by use of divergence and curl characteristics of the flow field.
• Exploit a physical characteristic descriptor of crowd motion (PCM).
• Calculate the conditional entropy of PCM distribution to perceiving the crowd changes.
摘要
•Explore the crowd state change detection problem for the dense crowd scenarios of wide-field of view in the public place.•Use a less computation required thermal diffusion processing and perspective transform processing to capture the macroscopic crowd motion information.•Model motion information by use of divergence and curl characteristics of the flow field.•Exploit a physical characteristic descriptor of crowd motion (PCM).•Calculate the conditional entropy of PCM distribution to perceiving the crowd changes.
论文关键词：Crowd state analysis,Physical characteristics,Temporal context of motion
论文评审过程：Received 9 August 2018, Revised 9 November 2018, Accepted 17 November 2018, Available online 19 November 2018, Version of Record 3 December 2018.
论文官网地址：https://doi.org/10.1016/j.patcog.2018.11.023
