Supervised Texture Segmentation: A Comparative Study

TL;DR

This study compares four feature extraction methods for texture segmentation, finding Gabor filters provide the best segmentation quality, while GLCM better localizes texture boundaries.

Contribution

It offers a comparative analysis of Gabor filters, GMRF, RLM, and GLCM for texture segmentation, highlighting their relative strengths.

Findings

Gabor filters achieved the highest segmentation quality.

GLCM provided superior boundary localization.

GMRF and RLM showed moderate performance.

Abstract

This paper aims to compare between four different types of feature extraction approaches in terms of texture segmentation. The feature extraction methods that were used for segmentation are Gabor filters (GF), Gaussian Markov random fields (GMRF), run-length matrix (RLM) and co-occurrence matrix (GLCM). It was shown that the GF performed best in terms of quality of segmentation while the GLCM localises the texture boundaries better as compared to the other methods.