Human Tracking Based on Mean Shift and Kalman Filter

In this paper we present a new method combining mean shift with Kalman filter for human tracking. Firstly, we use the mean shift algorithm based on color and texture features to calculate an accurate location in current frame. We select the HSV color space for calculating the histogram. Then Kalman filter is applied to predict the next initial searching location for mean shift iterations in the next frame. With this method, the target can be tracked successfully even when there is a large movement between two consecutively processed frames. Besides, this algorithm is also effective in the environment where the color distribution is extremely similar between the target and the background. In such an environment, the target can not be tracked correctly with the mean shift algorithm based on the histogram in RGB color space and the method of background subtraction will fail. Experimental results show that our algorithm is effective, robust and can be used for tracking in different scenes.