Algorithm 1Food plate detection

Input: an RGB image captured by Kinect Output: four corner points of the food plate 1.       Calculate the center point p of the captured image. The point p would locate inside the food plate. 2.       Draw a vertical ray along the y-axis with starting point p and upwards until it meets a red line. The red line L1 is selected as the upper edge of the plate. Calculate the red line’s slope and the angle θ between the line and x-axis. 3.       Draw a parallel rightwards ray of L1 with starting point p until it meets a red line and denoted as L2. 4.       Draw a parallel leftwards ray of L1 with starting point p until it meets a red line and denoted as L3. 5.       Draw a vertical downwards ray of L1 with starting point p until it meets a red line and denoted as L4. 6.       Determine the four corners from the extracted four lines L1 to L4 and redraw a rectangle (bounding box) and then output the four corner points.