Post-processing of Closed Contours to Obtain Inscribed K-Sided Polygons

This paper presents an algorithm for determining the maximum area or perimeter of a simple or convex polygon with k sides located within a region of interest. Inscribed polygons are used to detect, classify, and segment objects in images. Traditional algorithms generate basic shapes such as triangles, quadrilaterals, and pentagons. However, a user may be interested in obtaining polygons with the maximum area or perimeter, deciding whether they must be simple or convex, and even specifying the type of polygon needed (triangle, quadrilateral, etc.). The problem of determining the maximum area or perimeter of a simple k-sided polygon within a region of interest has not been solved in a generic way for a user-defined value of k. This paper presents a flexible algorithm that allows the user to determine which polygons should be calculated. The algorithm’s source code in C++, Java and Python is available in a GitHub repository to ensure its usability for the scientific community.