Visual Target Tracking Based on Compressive Feature Weighting

DOI

10.2991/iiicec-15.2015.153How to use a DOI?

Keywords

Visual Target Tracking; Compressive Features; Feature Weighting

Abstract

At present, most of the features used for tracking are high in dimension, and each dimension of the feature have the same weight, which will increase the amount of computation during tracking and may lead to a poor tracking performance. In this paper, we propose a visual target tracking algorithm based on compressive feature weighting. Unlike the existing methods, we extract compressive features of the target to reduce the amount of calculation, and each dimension of the compressive features is given a different weight by feature weighting method, then the weighted compressive features are used in a Bayes classifier to track the target. In addition, to reduce information loss, multiple compressive features are extracted, and a Kalman Filter is also used to improve the tracking performance. Experimental results demonstrate that the proposed tracking algorithm can achieve a very good performance in some challenging environments, even when the camera is moving.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Yinzhe Zhang
AU  - Guixi Liu
AU  - Hongyan Duan
PY  - 2015/03
DA  - 2015/03
TI  - Visual Target Tracking Based on Compressive Feature Weighting
BT  - Proceedings of the 2015 International Industrial Informatics and Computer Engineering Conference
PB  - Atlantis Press
SP  - 673
EP  - 676
SN  - 2352-538X
UR  - https://doi.org/10.2991/iiicec-15.2015.153
DO  - 10.2991/iiicec-15.2015.153
ID  - Zhang2015/03
ER  -