Video manifold modelling: finding the right parameter settings for anomaly detection

Authors:
Hanhe Lin;Jeremiah D. Deng;Brendon J. Woodford
Affiliations:
University of Otago, Dunedin, New Zealand;University of Otago, Dunedin, New Zealand;University of Otago, Dunedin, New Zealand
Venue:
Proceedings of the 27th Conference on Image and Vision Computing New Zealand
Year:
2012

Citing 7
Cited 0

Laplacian Eigenmaps for dimensionality reduction and data representation

Neural Computation
Image Spaces and Video Trajectories: Using Isomap to Explore Video Sequences

ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
Video event segmentation and visualisation in non-linear subspace

Pattern Recognition Letters
Event monitoring via local motion abnormality detection in non-linear subspace

Neurocomputing
Real-Time Abnormal Event Detection in Complicated Scenes

ICPR '10 Proceedings of the 2010 20th International Conference on Pattern Recognition
Learning video manifold for segmenting crowd events and abnormality detection

ACCV'10 Proceedings of the 10th Asian conference on Computer vision - Volume Part I
Learning and Matching of Dynamic Shape Manifolds for Human Action Recognition

IEEE Transactions on Image Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Using video manifold to analyze video scenes and detect possible anomaly has become a popular research topic in recent years. While a number of attempts have been proposed and reported promising outcomes, there is currently a lack of understanding about the parameter setting for various components in the algorithmic framework. In this paper we look at some key parameters, particularly the dimension of the video manifold, the embedding dimension of the video trajectory, and explore the plausibility of setting these parameters automatically using outcome of spectral clustering and fractal dimension analysis. Experiments are conducted using a benchmark dataset and the results are promising.