Skip to Main Content
Because when moving objects are tracked by mean-shift algorithm, the starting position of target in the current frame starts at the center of the target in the previous frame, so the accuracy of vehicle tracking is not too high. In order to overcome this shortcoming, the mask-based mean-shift algorithm is proposed in the paper. The accuracy of target tracking can be improved by forecasting its starting position. Experiments verify that the accuracy of vehicle tracking is improved by using the algorithm and that the problem of partial occlusion could be resolved.
Date of Conference: 18-20 Jan. 2009