Support Vector Machines Applied to 2-D Binary Image Euler Number Computation

Abstract

In this communication we explain how a Support Vector Machine (SVM) can be applied to compute the Euler number or Genus of a 2-D binary image. By taking into account the results provided by a mathematical formulation that is known producing exact results we derive two specialized SVM-based architectures, one useful for the 4-connected case and one useful for the 8-connected case. We validate the applicability of both architectures with two sets of images and compare their performance in terms of processing times with several standard methods reported in literature.