Appraisal of an enhanced particale filter for object tracking

  • Authors:

  • Howida A. Abd El-Halym;Imbaby I. Mahmoud;A. AbdelTawab;S. E.-D. Habib

  • Affiliations:

  • Nuclear Research Center, Atomic Energy Authority;Nuclear Research Center, Atomic Energy Authority;Misr University of Science and Technology;Faculty of Engineering, Cairo University

  • Venue:
  • ICIP'09 Proceedings of the 16th IEEE international conference on Image processing
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

An enhanced Particle Filter (PF) is introduced for object tracking. In this work, a new likelihood model is proposed. It depends on multiple of likelihood functions: position likelihood; gray level intensity likelihood; and similarity likelihood. Also, it combines information about the tracked object to get a robust and an accurate tracking performance. The proposed enhanced PF is implemented and evaluated. Its results are compared with a single likelihood function PF tracker, as well as, a correlation tracker and an edge tracker. The experimental results demonstrate the superior performance of the proposed tracker in terms of accuracy, robustness and occlusion compared with other methods.