All the examples in this chapter that deal with cameras return when there is a single failed or corrupted frame that leads to the detection of an empty frame. What type of modification is needed to allow a predefined number of retries before stopping the process?
How can we call the meanShift function to perform the Mean Shift algorithm with 10 iterations and an epsilon value of 0.5?
How can we visualize the hue histogram of the tracked object? Assume CamShift is used for tracking.
Set the process-noise covariance in the KalmanFilter class so that the filtered and measured values overlap. Assume only the process-noise covariance is set, of all the available matrices for KalmanFilter class-behavior control.
Let's assume that the Y position of the mouse on a window is used to describe the height of a filled rectangle that starts from the top-left corner of the window and has a width that equals the window width. Write a Kalman filter that can be used to correct the height of the rectangle (single value) and remove noise in the mouse movement that will cause a visually smooth resizing of the filled rectangle.
Create a BackgroundSubtractorMOG2 object to extract the foreground image contents while avoiding shadow changes.
Write a program to display the current (as opposed to sampled) background image using a background-segmentation algorithm.
