Superpixels-based Segmentation and Classification using Modified Fuzzy C-Means Clustering and Color Features

Abstract

This work provides a novel method of image analysis that combines color characteristics with Modified Fuzzy C-Means (MFCM) clustering by means of superpixels based segmentation and classification. The suggested technique improves the effectiveness and precision of image segmentation by utilizing the advantages of Superpixel representation. A hierarchical framework that minimizes computing complexity and maintains spatial coherence is used to construct superpixels. After that, the MFCM clustering technique is used to split superpixels more precisely, improving border adherence and adaptability while managing intricate image structures. Furthermore, color features are included to capture fine details in every cluster. The use of color information improves the suggested method’s ability to discriminate between objects that have similar textures but differing chromatic properties. The experimental findings demonstrate how well the suggested framework can handle problems pertaining to segmentation and classification tasks.