Mean shift tracking algorithm combined with Kalman Filter

This paper proposes a tracking algorithm that combines Mean Shift with Kalman Filter. Firstly, we use the Kalman filter to predict the target location. Secondly, the Mean Shift tracking algorithm is used to compute the target location with a linear weighted manner. The computed location is treated as a seed point. Finally, the Mean Shift searches target around seed point. Experimental results show that the algorithm can solve the target with suddenly velocity changes, and can more accurately predict the speed of the dynamic target to achieve an accurate tracking.