The Open Automation and Control Systems Journal

2015, 7 : 1022-1028
Published online 2015 September 10. DOI: 10.2174/18744443015070101022
Publisher ID: TOAUTOCJ-7-1022

Research on Mean-Shift Target Tracking Based on Image HOG Feature

Yang Xinfeng and Tian Di
School of Computer and Information Engineering, Nanyang Institute of Technology, Nanyang, 473004, P.R. China.

ABSTRACT

In this paper, mean shift algorithm based on HOG feature has been used to track the moving object. A feature hierarchical model has been constructed, which is coarser in the entire moving object layer. It is partly based on model components featuring accurate tracking of the moving object in different layers. The moving object is tracked using nuclear tracking with two layers which is more accurate than the moving object tracking with one layer information. Partially, it is based on the whole occlusion target location that can better be forecasted. The algorithm has a good tracking performance after the moving object appears again.

Keywords:

Hierarchical modeling, Histogram of oriented gradient, Mean shift, Moving object, Detection, Feature.