Detecting separation of moving objects based on non-parametric bayesian scheme for tracking by particle filter

Authors:
Yasuchika Takeda;Shinji Fukui;Yuji Iwahori;Robert J. Woodham
Affiliations:
Faculty of Engineering, Chubu University, Kasugai, Japan;Faculty of Education, Aichi University of Education, Kariya, Japan;Faculty of Engineering, Chubu University, Kasugai, Japan;Department of Conputer Science, University of British Columbia, Vancouver, B.C. Canada
Venue:
KES'11 Proceedings of the 15th international conference on Knowledge-based and intelligent information and engineering systems - Volume Part IV
Year:
2011

Citing 3
Cited 0

Deterministic annealing EM algorithm

Neural Networks
CONDENSATION—Conditional Density Propagation forVisual Tracking

International Journal of Computer Vision
MCMC-Based Particle Filtering for Tracking a Variable Number of Interacting Targets

IEEE Transactions on Pattern Analysis and Machine Intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper proposes a new approach for detecting separation of a group of moving objects tracked by the method based on the particle filter. When some objects with overlaps come into the area filmed with the camera, they are treated as one object and tracked by a group of particles. When the object group separates into each object or some groups, the method fails in tracking or there are some objects which are not tracked. The proposed method detects the separation of the object group by the non-parametric Bayesian scheme. The method can perform the whole process in shorter time than the previous method and remains the same performance.