Applications that interpret video data need to track objects as they move in a scene. Tracking methods that estimate the state trajectories of objects as they change over time (e.g. Kalman filter) have difficulty as the number of objects and clutter increase. We present an alternative, called statistical tracking, that is based on the concept of network tomography. A scene is modeled as a network of interconnected regions. Statistical tracking estimates the number of trips made from one region to another based on inter-region boundary traffic counts accumulated over time. It is not necessary to track an object through a scene, just to determine when an object crosses a boundary, something that is generally easier than estimating a continuous trajectory. In achieving this simplicity, statistical tracking gives up the ability to determine an object's state as the motion occurs. Instead, it determines mean traffic intensities based on statistics accumulated over a period of time. In spite of this limitation, there are several applications for which statistical tracking is useful. We demonstrate the application of the method to a large sample of video traffic surveillance data. The method does not require any data association which has some important implications concerning personal privacy.
[1]
Y. Vardi,et al.
Network Tomography: Estimating Source-Destination Traffic Intensities from Link Data
,
1996
.
[2]
Y. Bar-Shalom.
Tracking and data association
,
1988
.
[3]
Isaac Cohen,et al.
Detecting and Tracking Moving Objects in Video from an Airborne Observer
,
1998
.
[4]
C. Qian,et al.
Frame-rate Multi-body Tracking for Surveillance
,
1998
.
[5]
Ivana Mikic,et al.
Video Processing and Integration from Multiple Cameras
,
1998
.
[6]
David D. Sworder,et al.
Image fusion for tracking manoeuvring targets
,
1997,
Int. J. Syst. Sci..
[7]
Boris Polyak,et al.
The method of projections for finding the common point of convex sets
,
1967
.
[8]
D. Youla,et al.
Image Restoration by the Method of Convex Projections: Part 1ߞTheory
,
1982,
IEEE Transactions on Medical Imaging.
[9]
Arthur Gelb,et al.
Applied Optimal Estimation
,
1974
.
[10]
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).