Abstract: An improved Mean-Shift based tracking algorithm was proposed to solve the poor tracking ability problem in occlusions. Combinating multi-scale space theory, Kalman filter and occlusion, it can have a good tracking about scale using Kalman when target has been occluded. The experimental results show that the improved algorithm can select the proper size of the tracking window in the scenarios when the target has been occluded.

Key words: Kalman, Amount of information measure, McanShift, Occlusion