Material phase classification by means of Support Vector Machines

TL;DR

This paper introduces a Support Vector Machine-based method for classifying pixels in images of heterogeneous materials, improving accuracy over traditional thresholding methods, and discusses its impact on physical property calculations.

Contribution

The paper presents a novel SVM-based classification approach utilizing gradient features, outperforming Otsu's method in accuracy for material image analysis.

Findings

SVM achieved 77.6% accuracy, significantly higher than Otsu's 40.9%.

Improved classification enhances the reliability of physical property estimations.

Discussion shows better correlation function results with SVM classification.

Abstract

The pixel's classification of images obtained from random heterogeneous materials is a relevant step to compute their physical properties, like Effective Transport Coefficients (ETC), during a characterization process as stochastic reconstruction. A bad classification will impact on the computed properties; however, the literature on the topic discusses mainly the correlation functions or the properties formulae, giving little or no attention to the classification; authors mention either the use of a threshold or, in few cases, the use of Otsu's method. This paper presents a classification approach based on Support Vector Machines (SVM) and a comparison with the Otsu's-based approach, based on accuracy and precision. The data used for the SVM training are the key for a better classification; these data are the grayscale value, the magnitude and direction of pixels gradient. For instance, in the case study, the accuracy of the pixel's classification is 77.6% for the SVM method and 40.9% for Otsu's method. Finally, a discussion about the impact on the correlation functions is presented in order to show the benefits of the proposal.