Rectifying Reverse Polygonization of Digital Curves for Dominant Point Detection

A polygonal approximation technique using reverse polygonization is presented in this paper. This method rectifies the reverse polygonization in previous publications since it captures an efficient approximation independent to all initial points of curves. previous algorithms do not account for this, the algorithm presented in this article is more accurate than previous, specially when specified dominants for deleting are near to each other and follow a particular pattern. Our new approach starts from an initial set of dominant points i.e. break points and dominant points are deleted (one in each iteration) in a way that the maximal perpendicular distance of an approximating straight line from an original curve is minimized. The deletion is performed in a way that optimizes the results of approximation. Leftover points will be considered as dominant points that are not related to the starting point. Finally, a comparative study with former algorithms is provided which prove that under usage of this new technique better approximation results are obtainable.