Belongie and Malik [2], presented the "`Shape context"' feature descriptor describing shapes in a way that allows for
measuring shape similarity and the recovering of point correspondences. The basic idea of shape contexts is illustrated in Figure 1. A shape is represented by a discrete setPs = p1; :::; pn of n points sampled from the internal and
external contours on the shape. the n-1 vectors which are
the set of vectors originating from a point to all other sample
is a rich but too detailed description of the shape. Therefore, the authoress suggested to capture this information as
the distribution of the relative positions of the remaining
n-1 points in a spatial histogram, See Equation 1. This
histogram is defined to be the shape context of the origin
point pi.