ACLIME: Automatic cluster identification and merging

Clustering is an important phase in image and data analysis for assemblage of the items of almost similar types. Variety of real life application areas uses clustering methodology for result analysis. Among all, K-means algorithm is the most widely used unsupervised clustering approach as seen from past. However, for a particular domain specific problem the initial selection of K is still a great concern. In this paper a new methodology named ACLIME (Automatic Cluster Identification and Merging) is presented which is involuntarily found the optimal value of k for specific domain problems. The proposed approach is iterative and converges when an optimal solution is found out. ACLIME is based on the cluster performance measure analysis using intra and inter cluster similarity. Next, for result analysis proposed algorithm is tested for clustering medical image and scattered data points.