Texture Image Segmentation Using MATLAB with Gaussian Filtering and K-means Clustering

This project presents a MATLAB-based texture image segmentation program that employs a two-stage processing pipeline. First, we apply Gaussian filtering to smooth the input image, reducing noise while preserving important texture features through careful sigma parameter selection in the imgaussfilt() function. This preprocessing step enhances the subsequent clustering performance by creating more homogeneous regions. The core segmentation utilizes K-means clustering algorithm (implemented using kmeans() function) to partition the image into distinct texture regions based on pixel intensity and texture characteristics. The algorithm automatically determines optimal cluster centers through iterative optimization, with the cluster number parameter adjustable for different texture complexities. The segmented regions are then clearly delineated using boundary lines created through morphological operations and edge detection techniques. During extensive testing, our program demonstrated robust performance across various texture images, effectively handling different pattern complexities and scale variations. The implementation includes parameter optimization options and visualization tools for result analysis. We believe this solution contributes valuable methodology to texture analysis research and provides practical implementation reference through comprehensive documentation including test images, technical PDF reports, presentation slides, and detailed Word documentation that explain algorithmic approaches and usage scenarios.