Identification and Classification of Crowd Activities

作     者：Manar Elshahawy Ahmed O.Aseeri Shaker El-Sappagh Hassan Soliman Mohammed Elmogy Mervat Abu-Elkheir 

作者机构：Department of Information TechnologyFaculty of Computers and InformationMansoura UniversityEgypt Department of Computer ScienceCollege of Computer Engineering and SciencesPrince Sattam Bin Abdulaziz UniversityAl-Kharj11942Saudi Arabia Department of Information SystemsFaculty of Computers and Artificial IntelligenceBenha UniversityEgypt Faculty of Computer Science and EngineeringGalala University435611SuezEgypt Department of Computer ScienceFaculty of Media Engineering and TechnologyGerman UniversityEgypt 

出 版 物：《Computers, Materials & Continua》 (计算机、材料和连续体（英文）)

年 卷 期：2022年第72卷第7期

页      面：815-832页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

主　　题：Crowd analysis individual detection You Only Look Once(YOLO) multiple object tracking kalman filter pose estimation 

摘      要：The identification and classification of collective people’s activities are gaining momentum as significant themes in machine learning,with many potential applications *** need for representation of collective human behavior is especially crucial in applications such as assessing security conditions and preventing crowd *** paper investigates the capability of deep neural network(DNN)algorithms to achieve our carefully engineered pipeline for crowd *** includes three principal stages that cover crowd analysis ***,individual’s detection is represented using the You Only Look Once(YOLO)model for human detection and Kalman filter for multiple human tracking;Second,the density map and crowd counting of a certain location are generated using bounding boxes from a human detector;and Finally,in order to classify normal or abnormal crowds,individual activities are identified with pose *** proposed system successfully achieves designing an effective collective representation of the crowd given the individuals in addition to introducing a significant change of crowd in terms of activities *** results onMOT20 and SDHA datasets demonstrate that the proposed system is robust and *** framework achieves an improved performance of recognition and detection peoplewith a mean average precision of 99.0%,a real-time speed of 0.6ms non-maximumsuppression(NMS)per image for the SDHAdataset,and 95.3%mean average precision for MOT20 with 1.5ms NMS per image.