Two Algorithms for Shaped-Based Porosity Measures

Authors

Leen-Kiat Soh, University of NebraskaFollow

Date of this Version

2002

Document Type

Article

Comments

University of Nebraska-Lincoln, Computer Science and Engineering
Technical Report # TR-UNL-CSE-2002-0003

Abstract

In this paper, we present two algorithms for shape-based porosity measures. Specifically, we describe two innovative attributes called perimeter porosity and area porosity as feature descriptors. Perimeter porosity refers to how “porous” or “weak” an image object is, while area porosity determines the degree of “branchiness” of an object. We show the applicability of the two measures through a SAR sea ice image classification example.