The document investigates the performance of two cluster-based content-based image retrieval (CBIR) systems that fuse image visual features to enhance retrieval effectiveness. Through experimental comparisons with existing CBIR systems (UFM and CLUE), the authors demonstrate that their methods outperform the alternatives across various image resolutions. The analysis encompasses retrieval rates at different precision levels, establishing the advantages of the recommended systems in the context of unsupervised image retrieval.