Alireza Hakimi - International campus, Department of Computer Engineering, University of Isfahan, Isfahan, Iran
Peyman Adibi - Department of Computer Engineering, Faculty of Engineering, University of Isfahan, Isfahan, Iran

Clustering is an essential technique in some field of researches such as pattern recognition and machine learning. Based on variety of datasets and applications, a lot of clustering algorithms have been proposed. One of the differences between them is using distinct distance functions. Distance functions are the core of clustering algorithms and play the main role in organizing data into homogeneous groups. As a matter of fact, choosing the proper distance function is an open problem yet. In this paper, we propose a novel approach to Morphology Based Clustering (MBC). MBC employs dilation operator instead of any other distance function. Finally MBC was tested by two different datasets. Results show that MBC is able to find irregular shaped clusters and it is not sensitive to noise and outliers. The number of clusters and their specifications is alike after repetition of MBC algorithm in data sets

Clustering, Distance function, Dilation, mathematical morphology, Binary image