Large scale crowd analysis based on convolutional neural network》.pdfVIP

Pattern Recognition 48 (2015) 3016–3024 Contents lists available at ScienceDirect Pattern Recognition journal homepage: /locate/pr Large scale crowd analysis based on convolutional neural network Lijun Cao, Xu Zhang, Weiqiang Ren, Kaiqi Huang Center for Research on Intelligent Perception and Computing (CRIPC) National Laboratory of Pattern Recognition (NLPR) Institute of Automation, Chinese Academy of Sciences (CASIA), China a r t i c l e i n f o a b s t r a c t Article history: Nowadays crowd surveillance is an active area of research. Crowd surveillance is always affected by Received 8 October 2014 various conditions, such as different scenes, weather, or density of crowd, which restricts the real Received in revised form application. This paper proposes a convolutional neural network (CNN) based method to monitor the 6 March 2015 number of crowd flow, such as the number of entering or leaving people in high density crowd. It uses an Accepted 2 April 2015 indirect strategy of combining classification CNN with regression CNN, which is more robust than the Available online 15 April 2015 direct way. A large enough database is built with lots of real videos of public gates, and plenty of Keywords: experiments show that the proposed method performs well under various weather conditions no matter Crowd analysis either in daytime or at night. Flow co