To improve the effectiveness of pedestrian tracking, the histograms of oriented gradients (HOG) and color histogram characteristics are adopted to track pedestrian based on particle filter. Firstly, the pedestrian is detected using the HOG features to determine the initial target position. Then the target is tracked based on particle filter utilizing color histogram, during which the HOG is used to modify particle heavy weights and particle sampling. Experimental results verify the accurateness and efficiency of the proposed method.
Pedestrian Tracking with HOG and Color Histogram Features
Applied Mechanics and Materials ; 241-244 ; 498-501
13.12.2012
4 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Color-histogram-based tracking at 2000 fps
British Library Online Contents | 2012
|An EM-Like Algorithm for Color-Histogram-Based Object Tracking
British Library Conference Proceedings | 2004
|Efficient Color Histogram Indexing
British Library Conference Proceedings | 1994
|Efficient color histogram indexing
IEEE | 1994
|