Application of the self-organising map to trajectory classification

Owens, Jonathan and Hunter, Andrew (2000) Application of the self-organising map to trajectory classification. In: IEE Visual Surveillance Workshop , 1 July 2000, Dublin, Ireland.

Documents
Application of the self-organising map to trajectory classification
[img]
[Download]
Request a copy
Application of the self-organising map to trajectory classification
Publisher version
[img]
[Download]
[img] PDF
IEEE.PDF
Restricted to Repository staff only

175kB
[img]
Preview
PDF
06980077.pdf

923kB

Full text URL: http://doi.ieeecomputersociety.org/10.1109/VS.2000...

Abstract

This paper presents an approach to the problem of automatically classifying events detected by video surveillance systems; specifically, of detecting unusual or suspicious movements. Approaches to this problem typically involve building complex 3D-models in real-world coordinates
to provide trajectory information for the classifier. In this paper we show that analysis of trajectories may be carried out in a model-free fashion, using self-organising
feature map neural networks to learn the characteristics of normal trajectories, and to detect novel ones. Trajectories are represented using positional and first and second order motion information, with moving-average smoothing. This allows novelty detection to be applied on a point-by-point basis in real time, and permits both instantaneous motion and whole trajectory motion to be subjected to novelty detection.

Item Type:Conference or Workshop Item (Paper)
Additional Information:This paper presents an approach to the problem of automatically classifying events detected by video surveillance systems; specifically, of detecting unusual or suspicious movements. Approaches to this problem typically involve building complex 3D-models in real-world coordinates to provide trajectory information for the classifier. In this paper we show that analysis of trajectories may be carried out in a model-free fashion, using self-organising feature map neural networks to learn the characteristics of normal trajectories, and to detect novel ones. Trajectories are represented using positional and first and second order motion information, with moving-average smoothing. This allows novelty detection to be applied on a point-by-point basis in real time, and permits both instantaneous motion and whole trajectory motion to be subjected to novelty detection.
Keywords:Trajectory classification, self-organising map, surveillance
Subjects:G Mathematical and Computer Sciences > G760 Machine Learning
Divisions:College of Science > School of Computer Science
ID Code:1906
Deposited By: Tammie Farley
Deposited On:25 Jun 2009 14:52
Last Modified:13 Mar 2013 08:32

Repository Staff Only: item control page