Integration of projection histograms and linear prediction for object tracking

Abstract

In this paper, a real-time, efficient and robust visual tracking system for a single moving object using image sequences captured by a stationary camera is presented. The proposed tracking algorithm integrates the Projection Histograms technique with the Linear Prediction method in order to achieve a faster tracking speed. The Projection Histograms technique is applied to obtain the actual location of the tracked target, whereas the Linear Prediction method is incorporated in the proposed tracking algorithm to predict the location of the moving object in the next frame based on its several past centroid measurements. The Projection Histograms technique coupled with a second order Linear Prediction method has enabled the proposed algorithm to accurately track a single moving object. The potential applicability and efficiency of the proposed tracking algorithm has been validated by good tracking results on several experimental image sequences.