Adaptive Mean Shift and Particle Filter Tracking Method Based on Joint Feature

摘要

In the object tracking area, both particle filter and mean shift algorithm have proven successful approaches. However, both of them have notable weakness. In this paper, we present a new algorithm which combined the two algorithms to track the target. First, the mean shift algorithm is employed to search an object candidate near the target state. Then, if the candidate is good enough, it will be used to adapt the particle filter parameters, including the number of particle filter, and etc. Finally, the particle filter will estimate the target state based on these new parameters. Further, the paper introduces the color-texture combined feature instead of color feature.