Object Tracking and Behavior Recognition Based on Gray Prediction

DOI

10.2991/iiicec-15.2015.106How to use a DOI?

Keywords

Object Detection; Object Tracking; Gray Prediction Model; Behavior Recognition

Abstract

For existing problems in visual tracking and behavior recognition, we propose a novel method to track and recognize based on gray prediction in this paper. We firstly use the background subtraction method to detect moving target and use cross entropy method to process image binaryzation. After that, the morphology filter is used to eliminate the noise, in this way we extract the template whose size is determined by the contour segmentation. The improved gray prediction employs GM(1,1) model to reduce the prediction scope, such that human body’s motion trajectory can be recognized. Through the tracking and recording of human moving trajectory, we can identify moving human behavior, like jumping, tumble and squat. Experimental results prove that the proposed algorithm can recognize jumping, tumble, squat and other common human motion behavior correctly as well as very robust.

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  - Fang Zhang
PY  - 2015/03
DA  - 2015/03
TI  - Object Tracking and Behavior Recognition Based on Gray Prediction
BT  - Proceedings of the 2015 International Industrial Informatics and Computer Engineering Conference
PB  - Atlantis Press
SP  - 462
EP  - 465
SN  - 2352-538X
UR  - https://doi.org/10.2991/iiicec-15.2015.106
DO  - 10.2991/iiicec-15.2015.106
ID  - Zhang2015/03
ER  -