Introduction - If you have any usage issues, please Google them yourself
The first just does detection by background subtraction. This can be considered as the ground truth.
The second feeds the detection output into a Kalman filter. The predicted position from the kalman filter (red) is compared against the actual ground truth position (green).