Application of optimum-path forest classifier for synthetic material porosity segmentation

Abstract

This paper presents a new application and evaluation of the Optimum-Path Forest (OPF) classifier to accomplish synthetic material porosity segmentation and quantification obtained from optical microscopic images. Sample images of a synthetic material were analyzed and the quality of the results was confirmed by human visual analysis. Additionally, the OPF results were compared against two different Support Vector Machines approaches, confirming the OPF superior fast and reliable qualities for this analysis purpose. Thus, the Optimum-Path Forest classier demonstrated to be a valid and adequate tool for microstructure characterization through porosity segmentation and quantification using microscopic images, manly due its fast, efficient and reliable manner.