Efficient Multi-Target Human Motion Tracking

被引：0

作者：

Abhyankar, Aditya ^{[1
]}

机构：

[1] Clarkson Univ, ECE, Potsdam, NY 13676 USA

来源：

IACSIT-SC 2009: INTERNATIONAL ASSOCIATION OF COMPUTER SCIENCE AND INFORMATION TECHNOLOGY - SPRING CONFERENCE | 2009年

关键词：

Motion Tracking; Multiple Targets; Particle filtering; Markov Chains; Monte Carlo method;

D O I：

10.1109/IACSIT-SC.2009.100

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

Effective human motion tracking is necessary for all surveillance based biometric applications. Human motion being highly non-linear and non-modular is always difficult to track. To make it more challenging, in real scenarios, there are always multiple targets to be tracked. In this work particle filter based multiple target tracking system is proposed. For performing actual tracking Markov Chain Monte Carlo methods are used. The system was tested using moviedb and labdb databases and accuracy of 91.7% was obtained.

引用

页码：571 / 575

页数：5