Given a binary image I of a connected component (CC),
the process starts with calculating the centroid C and diameter D of the image I. The values of C and D are used
to determine and draw a set of rings centered by C with
different radius values derived from the diameter D. We
treat the rings as lying on different heights above the given
shape where larger rings overlay closer to the shape and the
smallest ring is placed on the top, see Figure 1. In the next
step, we treat each ring as a set of k points taken uniformly
distant from each other. Each point in each ring serves as
an upper view point watching the pixels in the shape. When
the MAD descriptor is used as a boundary based descriptor,
we derive their order as they are extracted from the contour
consistently. When the descriptor uses the image as an edge-map/multicomponent image, the pixels order is derived by
sorting the vectors connecting pixels to the centroid using
angles to the X - Axis.