Video object tracking based on position prediction guide CAMSHIFT

In this article, a novel algorithm-position prediction guide continuously adaptive mean shift procedure (CAMSHIFT)-is proposed for tracking objects in video sequences. CAMSHIFT is incorporated with the position prediction algorithm makes the scale adaptation of CAMSHIFT is improved, tracking efficiency increased and the computation complexity is reduced. Preliminary experimental results show the algorithm performs well. Meanwhile, the algorithm success in indicating new objects appeared; objects disappeared; objects collisions and the consistency of the objects in consecutive sequences.