A Hierarchical Clustering Based on Overlap Similarity Measure

Selecting the appropriate number of clusters and distinguishing partially overlapping and irregular data are two important problems in clustering. Hierarchical clustering provides a good solution to them. Similarity measure is the key of controlling the iterative process of hierarchical clustering. In this paper, we give a definition of overlap similarity measure and proposed a hierarchical clustering algorithm based on it without specified number of clusters in advance, whose appropriate value can be decided in the iterative process. The algorithm stops clustering according to the overlap similarity between clusters. Clustering analysis is a useful approach to unsupervised image segmentation. After discussing some related topics, we applied it to synthetic and real image segmentation to evaluate the performance of the clustering algorithm and compared it with other algorithms. Moreover, we estimated parameters of the algorithm in image segmentation. Experimental results show that this approach can be effectively applied to image segmentation.