Counting people from multiple cameras

We are interested in the content analysis of video from a collection of spatially distant cameras viewing a single environment. We address the task of counting the number of different people walking through such an environment, which requires that the system can identify which observations from different cameras show the same person. Our system achieves this by combining visual appearance matching with mutual content constraints between the cameras. We present results from a system with four very different camera views that counts people walking through and around a research lab.

[1]  François Bourgeois,et al.  An extension of the Munkres algorithm for the assignment problem to rectangular matrices , 1971, CACM.

[2]  Ronald A. Howard,et al.  Dynamic Probabilistic Systems , 1971 .

[3]  W. Eric L. Grimson,et al.  Using adaptive tracking to classify and monitor activities in a site , 1998, Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231).

[4]  Stuart J. Russell,et al.  Object Identification: A Bayesian Analysis with Application to Traffic Surveillance , 1998, Artif. Intell..

[5]  Aubrey B. Poore,et al.  Multidimensional assignment formulation of data association problems arising from multitarget and multisensor tracking , 1994, Comput. Optim. Appl..

[6]  Jeffrey E. Boyd,et al.  MPI-Video infrastructure for dynamic environments , 1998, Proceedings. IEEE International Conference on Multimedia Computing and Systems (Cat. No.98TB100241).

[7]  Olga Veksler,et al.  A Variable Window Approach to Early Vision , 1998, IEEE Trans. Pattern Anal. Mach. Intell..