一种Mean Shift视频目标跟踪算法的改进

Mean Shift tracking algorithm is a robust fast feature matching algorithm, which lacks of the necessary target model update, so the overall model update algorithm is proposed, but this method still can not efficiently hand the appearance changes on the scene. Therefore, this article presents a selective model update algorithm applied to Mean Shift tracking algorithm, in which each component of the feature model is regarded as a separate component, according to the value of the matching contribution for each component, selecting the component of current frame that needs to be updated. In the end, the improved algorithm is realized on VC++ 6.0 software platform. The experimental results show that the newly proposed algorithm is better for tracking robustness.