ABSTRACT
Contents 16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358 16.2 Related Work and Our Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
16.2.1 Comparison for Existing Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359 16.2.2 Our Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
16.3 Technical Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362 16.3.1 Motivation and Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362 16.3.2 Game-Theoretic Framework. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
16.3.2.1 Computation of Utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363 16.3.2.2 Bargaining among Cameras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365 16.3.2.3 Criteria for Camera Assignment and Handoff . . . . . . . . . . . . . . . . . . . . . . 366
16.3.3 Theoretical Comparison with Two Non-Game-Theoretic Approaches . . . . . . . 367 16.3.3.1 Descriptions of the Key Ideas of Selected Approaches . . . . . . . . . . . . . . 367 16.3.3.2 Pros and Cons Comparison of the Selected Approaches . . . . . . . . . . . . 369
16.4 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369 16.4.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370 16.4.2 Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370 16.4.3 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371 16.4.4 Experimental Results and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
16.5 Conclusions and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
Due to the broad coverage of an environment and the possibility of coordination among different cameras, video sensor networks have attracted much interest in recent years. Although the field of view (FOV) of a single camera is limited and cameras may have overlapping or nonoverlapping FOVs, seamless tracking of moving objects can be achieved by exploiting the handoff capability of multiple cameras. In this chapter, we will provide a new perspective to the camera selection and handoff problem that is based on game theory. In our work, game theory is used for multicamera multi-person seamless tracking based on a set of user-supplied criteria in a network of video cameras for surveillance and monitoring. The bargaining mechanism is considered for collaborations as well as for resolving conflicts among the available cameras. Camera utilities and person utilities are computed based on a set of criteria. They are used in the process of developing the bargaining mechanisms. The merit of our approach is that it is independent of the topology of how the cameras are placed in the network. When multiple cameras are used for tracking and where multiple cameras can “see” the same object, we are able to choose the “best” camera based on multiple criteria that are selected a priori. The algorithm can automatically provide an optimal as well as stable solution of the camera assignment quickly. The detailed camera calibration or 3D scene understanding is not needed in our approach. Experiments for multi-camera multi-person tracking are provided to corroborate the proposed approach. We also provide a comprehensive comparison of our work and some non-game-theoretic approaches, both theoretically and experimentally.
