Fast and effective color-based object tracking by boosted color distribution

In this paper, we propose a novel tracking algorithm, boosted color distribution (BCD), for tracking color objects. There exist three contributions in this paper. First, we propose a novel online gentle boost (OGB) algorithm for online learning. The essential idea of OGB is composed of two aspects: online updating candidate weak classifiers, and then choosing and combining them in a boosting way. Second, we design a novel weak classifier, log color feature distribution ratio, which focuses on the difference of color distributions rather than individual samples and provides a simple yet effective manner of mining color features for object tracking. Third, by combining our OGB algorithm and our log color features, we develop a fast yet effective color-based object tracking algorithm. Numerous experiments demonstrate that our tracking algorithm is better than or not worse than some state-of-the-art tracking algorithms on some public sequences.Overall, this paper presents a novel BCD algorithm for color object tracking that achieves good results at a fast speed.