Novelty Detection in Video Surveillance Using Hierarchical Neural Networks

Authors:
Jonathan Owens;Andrew Hunter;Eric Fletcher
Affiliations:
-;-;-
Venue:
ICANN '02 Proceedings of the International Conference on Artificial Neural Networks
Year:
2002

Citing 8
Cited 5

Learning invariance from transformation sequences

Neural Computation
Learning the distribution of object trajectories for event recognition

BMVC '95 Proceedings of the 6th British conference on Machine vision (Vol. 2)
Position invariant recognition in the visual system with cluttered environments

Neural Networks
Learning Patterns of Activity Using Real-Time Tracking

IEEE Transactions on Pattern Analysis and Machine Intelligence
Multimedia Video-Based Surveillance Systems: Requirements, Issues and Solutions

Multimedia Video-Based Surveillance Systems: Requirements, Issues and Solutions
A hybrid learning network for shift-invariant recognition

Neural Networks
Application of the Self-Organizing Map to Trajectory Classification

VS '00 Proceedings of the Third IEEE International Workshop on Visual Surveillance (VS'2000)
A topological and temporal correlator network for spatiotemporal pattern learning, recognition, and recall

IEEE Transactions on Neural Networks

Multiple object tracking using a neural cost function

Image and Vision Computing
FPGA-Based Anomalous Trajectory Detection Using SOFM

ARC '09 Proceedings of the 5th International Workshop on Reconfigurable Computing: Architectures, Tools and Applications
Motion-based behaviour learning, profiling and classification in the presence of anomalies

Pattern Recognition
Behavioral analysis of mobile robot trajectories using a point distribution model

SAB'06 Proceedings of the 9th international conference on From Animals to Animats: simulation of Adaptive Behavior
L1 norm based KPCA for novelty detection

Pattern Recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

A hierarchical self-organising neural network is described for the detection of unusual pedestrian behaviour in video-based surveillance systems. The system is trained on a normal data set, with no prior information about the scene under surveillance, thereby requiring minimal user input. Nodes use a trace activation rule and feedforward connections that are modified so higher layer nodes are sensitive to trajectory segments traced across the previous layer. Top layer nodes have binary lateral connections and corresponding "novelty accumulator" nodes. Lateral connections are set between co-occurring nodes, generating a signal to prevent accumulation of the novelty measure along normal sequences. In abnormal sequences the novelty accumulator nodes are allowed to increase their activity, generating an alarm state.