Learning Semantic Scene Models From Observing Activity in Visual Surveillance

Abstract

This paper considers the problem of automatically learning an activity-based semantic scene model from a stream of video data. A scene model is proposed that labels regions according to an identifiable activity in each region, such as entry/exit zones, junctions, paths, and stop zones. We present several unsupervised methods that learn these scene elements and present results that show the efficiency of our approach. Finally, we describe how the models can be used to support the interpretation of moving objects in a visual surveillance environment.

Keywords

This publication has 14 references indexed in Scilit:

A hierarchical database for visual surveillance applications
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2005
Multi view image surveillance and tracking
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2003
Estimating Tracking Sources and Sinks
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2003
Using adaptive tracking to classify and monitor activities in a site
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2002
Path detection in video surveillance
Image and Vision Computing, 2002
Spatial and Probabilistic Modelling of Pedestrian Behaviour
Published by British Machine Vision Association and Society for Pattern Recognition ,2002
Finding Paths in Video Sequences
Published by British Machine Vision Association and Society for Pattern Recognition ,2001
W/sup 4/: real-time surveillance of people and their activities
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2000
Learning the distribution of object trajectories for event recognition
Image and Vision Computing, 1996
Understanding object motion: Recognition, learning and spatiotemporal reasoning
Robotics and Autonomous Systems, 1991