Integrating Image Visibility Graph and Topological Data Analysis for Enhanced Texture Classification

In this study, a method for classifying textures based on image visibility graphs and topological data analysis is given. To improve texture classification result, we propose a new method that uses topological data analysis together with image visibility graphs. The present study involves the analysis of the degree distribution obtained from the image visibility graph along with the extraction of seven distinct topological features that are subsequently utilized for classification purposes. The proposed approach has been evaluated on established image texture datasets, such as the Salzburg texture image dataset. The results show an improvement in performance, suggesting the possibility of integrating graph-based techniques and topological characteristics in the process of texture classification.