Share Email Print

Proceedings Paper

Adaptive particle filter for robust visual tracking
Author(s): Jianghua Dai; Shengsheng Yu; Xiaoping Chen; Jinhai Xiang
Format Member Price Non-Member Price
PDF $17.00 $21.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Object tracking plays a key role in the field of computer vision. Particle filter has been widely used for visual tracking under nonlinear and/or non-Gaussian circumstances. In particle filter, the state transition model for predicting the next location of tracked object assumes the object motion is invariable, which cannot well approximate the varying dynamics of the motion changes. In addition, the state estimate calculated by the mean of all the weighted particles is coarse or inaccurate due to various noise disturbances. Both these two factors may degrade tracking performance greatly. In this work, an adaptive particle filter (APF) with a velocity-updating based transition model (VTM) and an adaptive state estimate approach (ASEA) is proposed to improve object tracking. In APF, the motion velocity embedded into the state transition model is updated continuously by a recursive equation, and the state estimate is obtained adaptively according to the state posterior distribution. The experiment results show that the APF can increase the tracking accuracy and efficiency in complex environments.

Paper Details

Date Published: 30 October 2009
PDF: 5 pages
Proc. SPIE 7495, MIPPR 2009: Automatic Target Recognition and Image Analysis, 74954O (30 October 2009); doi: 10.1117/12.833900
Show Author Affiliations
Jianghua Dai, Huazhong Univ. of Science and Technology (China)
Shengsheng Yu, Huazhong Univ. of Science and Technology (China)
Xiaoping Chen, Huazhong Univ. of Science and Technology (China)
Jinhai Xiang, Huazhong Univ. of Science and Technology (China)

Published in SPIE Proceedings Vol. 7495:
MIPPR 2009: Automatic Target Recognition and Image Analysis
Tianxu Zhang; Bruce Hirsch; Zhiguo Cao; Hanqing Lu, Editor(s)

© SPIE. Terms of Use
Back to Top