Presentation 2006-03-15
Adaptive Mixing of Proposal Distributions of Hierarchical Particle Filtering and its Application to Real-time Pose Estimation of Rigid Object
Takashi BANDO, Tomohiro SHIBATA, Mikio SHIMIZU, Shin ISHII,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) One of the major drawbacks of Particle Filters (PFs) is that a large number of particles are generally required for accurate estimation of state variables lying in a high dimensional space, which is time-consuming. In many applications, the high-dimensional state variables can be hierarchically modeled and this study presents an adaptive proposal distribution which is a mixture of prediction densities computed in individual layers, and its mixture ratio corresponding to the reliability of each layer is determined by means of an on-line EM algorithm. The adaptive proposal distribution enables robust and accurate estimation with a few numbers of particles, which is demonstrated by computer simulations of pose estimation of a rigid object as well as real experiments of driver's head pose estimation in a real car.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) particle filter / hierarchy / real-time processing / pose estimation / on-line EM algorithm
Paper # NC2005-119
Date of Issue

Conference Information
Committee NC
Conference Date 2006/3/8(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Adaptive Mixing of Proposal Distributions of Hierarchical Particle Filtering and its Application to Real-time Pose Estimation of Rigid Object
Sub Title (in English)
Keyword(1) particle filter
Keyword(2) hierarchy
Keyword(3) real-time processing
Keyword(4) pose estimation
Keyword(5) on-line EM algorithm
1st Author's Name Takashi BANDO
1st Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology()
2nd Author's Name Tomohiro SHIBATA
2nd Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology
3rd Author's Name Mikio SHIMIZU
3rd Author's Affiliation Vehicle Integrated Systems R & D Dept.
4th Author's Name Shin ISHII
4th Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology
Date 2006-03-15
Paper # NC2005-119
Volume (vol) vol.105
Number (no) 657
Page pp.pp.-
#Pages 6
Date of Issue