Comparative Analysis of Clustering Algorithms Based on Pixel Image Processing Using the Ssim Metric and Visual Analysis

The paper considers the problem of clustering pixels of a color raster image. The task is to compare the effectiveness of three different clustering methods: k-means, DBSCAN, agglomerative clustering. The k-means and agglomerative clustering algorithms consider different numbers of clusters: 2, 5, 10, 15, 20. The effectiveness is assessed using the SSIM metric and visual analysis of the resulting images.