ABSTRACT
In this chapter we present an algorithm [124] for polygonal approximation of digital curve which is based on discrete curvature measure. We propose that one can use k-cosine itself to determine the region of support without using any input parameter. A new measure of discrete curvature based on k-cosine called smoothed k-cosine is introduced. The local maxima and minima of smoothed k-cosine are located. We call these points as significant points. The adjacent significant points are joined to determine the polygon.
