Class Average Clustering Method for Image Recognition

This MATLAB-based GUI program implements class average clustering methodology for recognizing different ground features in various image types, with specialized focus on remote sensing imagery. The system architecture comprises three core components: a main program handling the clustering workflow, a FIG file defining the graphical interface layout, and three distinct distance calculation functions (Euclidean, Manhattan, and customized metrics) selectable through interactive parameters. Key implementation features include an optimized clustering algorithm that computes inter-cluster distances by averaging pairwise distances between all member points, providing balanced separation of heterogeneous image regions. The GUI interface enables parameter configuration for distance function selection, cluster number specification, and real-time visualization of segmentation results. This integrated approach enhances image processing accuracy and operational efficiency through intuitive controls and algorithm customization capabilities suitable for multi-spectral image analysis applications.