Efficient Fast Object-Tracking Scheme Based on Motion-vector-located Pattern Match

In the process of object tracking, the major problem is how to mark the tracking box of the object. Moreover, multi-objects tracking is also difficult. This paper proposed and efficient fast object-tracking scheme based on motion-vector-located pattern match, which adopts motion vector of Mpeg2 to mark the moving targets in static video in order to mark and locate the targets automatically and quickly. Then, extract multi-dimensional characteristics from the initial targets taken by motion vector and make the model. Then accurately identifies the particles of larger weight and combines with inertia factor of velocity through matching the original data and the observations of particle filter. The matching of the characteristics of the new particle and the original one is more accurate and faster because of adopting the method of pattern classifying. The experiments show that the algorithm had good tracking performance and strong robustness.