We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Remarkable local resampling based on particle filter for visual tracking.
- Authors
Zhao, Zhiqiang; Wang, Tianjiang; Liu, Fang; Choe, Gwangmin; Yuan, Caihong; Cui, Zongmin
- Abstract
Generally, particle filters need a large number of particles to approximate the posterior for the purpose of ideal effect. Previous methods extract remarkable particles from the particles at time t-1 by nonlinear function. Those methods use the remarkable particles to reduce the number of particles and improve the accuracy of particle filter. However, the nonlinear function extracts the remarkable particles, which will weaken or even ignore useful remarkable local particles. Thus this paper presents a new resampling scheme to extract remarkable local particles. We propose a weight threshold and a distance threshold to extract remarkable local particles from particles at time t-1. Meanwhile, we use these remarkable local particles to track the target analytically. Besides, we propose a global transition model to improve the accuracy of the particle filter. Based on remarkable local resampling scheme and the global transition model, we propose a new framework of particle filter. Finally, experiments show that our framework has higher efficiency than previous methods in the case of fewer particles.
- Subjects
TRACKING &; trailing; RESAMPLING (Statistics); NONPARAMETRIC statistics; MONTE Carlo method; NONLINEAR functions
- Publication
Multimedia Tools & Applications, 2017, Vol 76, Issue 1, p835
- ISSN
1380-7501
- Publication type
Article
- DOI
10.1007/s11042-015-3075-6