A Bayesian approach to joint tracking and identification of geometric shapes in video sequences

Authors:
Pierre Minvielle;Arnaud Doucet;Alan Marrs;Simon Maskell
Affiliations:
CEA, DAM, CESTA, F-33114 Le Barp, France;The Institute of Statistical Mathematics, Tokyo, Japan;QinetiQ Ltd., St. Andrews Road, Malvern, United Kingdom;QinetiQ Ltd., St. Andrews Road, Malvern, United Kingdom
Venue:
Image and Vision Computing
Year:
2010

Citing 8
Cited 0

On sequential Monte Carlo sampling methods for Bayesian filtering

Statistics and Computing
Annealed importance sampling

Statistics and Computing
Joint tracking of manoeuvring targets and classification of their manoeuvrability

EURASIP Journal on Applied Signal Processing
A survey of convergence results on particle filtering methods forpractitioners

IEEE Transactions on Signal Processing
A tutorial on particle filters for online nonlinear/non-GaussianBayesian tracking

IEEE Transactions on Signal Processing
Particle filters for state-space models with the presence ofunknown static parameters

IEEE Transactions on Signal Processing
A particle algorithm for sequential Bayesian parameter estimationand model selection

IEEE Transactions on Signal Processing
A gaussian sum approach to the multi-target identification-tracking problem

Automatica (Journal of IFAC)

Quantified Score

Hi-index	0.00

Visualization

Abstract

A Bayesian approach is proposed for joint tracking and identification. These two problems are often addressed independently in the literature, leading to suboptimal performance. In a Bayesian approach, a prior distribution is set on both the hypothesis space and the associated parameter space. Although this is straightforward from a conceptual viewpoint, it is typically impossible to perform inference in closed-form. We discuss an advanced particle filtering approach to solve this computational problem and apply this algorithm to joint tracking and identification of geometric forms in video sequences.