| EPSRC Reference: |
GR/K08772/01 |
| Title: |
BEHAVIOURAL ANALYSIS FOR VISUAL SURVEILLANCE USING BAYESIAN NETWORKS |
| Principal Investigator: |
Buxton, Professor H |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Sch of Engineering and Informatics |
| Organisation: |
University of Sussex |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 October 1994 |
Ends: |
30 April 1998 |
Value (£): |
162,194
|
| EPSRC Research Topic Classifications: |
|
| EPSRC Industrial Sector Classifications: |
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
To develop a computational framework for advanced surveillance incorporating information integration to give robust, timely behavioural analysis of video sequences of indoor and outdoor scenes. To develop Bayesian Belief Networks (BBNs) to model the causal relationships and task structure in the evaluation of visual behaviour under attentional control. To develop the context-dependent (deictic) representations of behaviour together with the analogical spatial representation of scenes for situated analysis. To develop adaptive techniques for the BBNs to allow the learning of behavioural invariances as pre-attentive cues for more flexible, automated analysis.Progress:The robust and reliable analysis of image sequences is essential for many new vision applications. This is particularly true in the field of visual surveillance. We are developing a systematic methodology for the design, integration and implementation of such systems using Bayesian networks. The key advantages of this approach are:1) Systematic handling of uncertainty and incompleteness in the data;2) fast, dynamic updating of the parameter network for context-based processing; and3) clear mapping of contextual knowledge onto the computational framework to constrain the interpretations. A SUN SPARC 20 2 processor workstation has been purchased and is performing at a high level with our initial behavioural analysis scheme using established data. Work is in progress to link model-based tracking software from Reading University to our higher-level analysis system. New data generated from the extended systems will be analysed using the fixed set of key features and compared with an adaptive scheme for finding key features for a set of simple behaviours.PublicationsH. Buxton and S.G. Gong. Advanced Visual Surveillance .Technical report, School of Cognitive and Computing Sciences, University of Sussex, 1993. CSRP 293 and submitted to special issue of AI Journal.S.G. Gong and H. Buxton. On the Expectations of Moving Objects .In European Conference on Artificial Intelligence, Vienna, Austria, 1992.R. Howarth. Spatial Representation, Reasoning and Control for a Surveillance System .PhD thesis, Department of Computer Sciences, QMW, University of London, 1994.R. Howarth and H. Buxton. An analogical representation of space and time .Image and Vision Computing, 10(7), 1992.R. Howarth and H. Buxton. Selective Attention in Dynamic Vision .In International Joint Conference on Artificial Intelligence, Chambery, France, 1993.
|
|
Key Findings |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Potential use in non-academic contexts |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Impacts |
|
| Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
| Summary |
|
| Date Materialised |
|
|
|
|
Sectors submitted by the Researcher |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
| Project URL: |
|
| Further Information: |
|
| Organisation Website: |
http://www.sussex.ac.uk |