Real-Time Distributed Multi-object Tracking in a PTZ Camera Network Ayesha Choudhary 1(B ) , Shubham Sharma 2 , Indu Sreedevi 2 , and Santanu Chaudhury 3 1 School of Computer and System Sciences, Jawaharlal Nehru University, New Delhi, India ayeshac@mail.jnu.ac.in 2 Department of Electronics and Communication Engineering, Delhi Technological University, New Delhi, India shubh2494@gmail.com, s.indu@rediffmail.com 3 Department of Electrical Engineering, Indian Institute of Technology Delhi, New Delhi, India santanuc@ee.iitd.ac.in Abstract. A visual surveillance system should have the ability to view an object of interest at a certain size so that important information related to that object can be collected and analyzed as the object moves in the area observed by multiple cameras. In this paper, we propose a novel framework for real-time, distributed, multi-object tracking in a PTZ camera network with this capability. In our framework, the user is provided a tool to mark an object of interest such that the object is tracked at a certain size as it moves in the view of various cameras across space and time. The pan, tilt and zoom capabilities of the PTZ cameras are leveraged upon to ensure that the object of interest remains within the predefined size range as it is seamlessly tracked in the PTZ camera network. In our distributed system, each camera tracks the objects in its view using particle filter tracking and multi-layered belief propagation is used for seamlessly tracking objects across cameras. Keywords: Distributed multi-camera tracking · Real-time tracking · PTZ camera network · Collaborative multi-object tracking · Belief propagation 1 Introduction A real-time video surveillance system consisting of a PTZ (pan, tilt, zoom) cam- era network requires seamless tracking of multiple objects in the scene. Moreover, particular objects of interest, such as suspects, may be required to be tracked at a certain dimension in each frame so that important information related to that object is continuously retained. In general, it is possible that the object of interest can become so small that a lot of information about the object is lost. On the other hand, the object of interest can come so close to a camera that the c  Springer International Publishing Switzerland 2015 M. Kryszkiewicz et al. (Eds.): PReMI 2015, LNCS 9124, pp. 183–192, 2015. DOI: 10.1007/978-3-319-19941-2 18