Optimal cluster numbers of unsupervised classification in Minkowski spaces

Unsupervised classification needs no prior knowledge of interested region and has less human interference that is suitable for processing messes of digital images acquired from earth observation satellites. A typical sequence for unsupervised classification might begin with the analyst specifying minimum and maximum numbers of categories to be generated by the classification algorithm. Usually, a larger cluster number is assigned in advance for the initial clustering. Then, several classified clusters are merged into a larger group so that the number of clusters can be reduced. The selection of an appropriate cluster number is up to the analyst's experience. This paper illustrates the convergence of cluster-center optimization and the optimal cluster number varying with the dimension in Minkowski space. One of the simplest and most popular is Euclidean distance while Minkowski distance is measured in two dimensions. The experiment was executed on several digital images by a classifier employing Genetic Algorithm as a searching tool and Davies-Bouldin index as a cluster validity measure. The results show that a smaller dimension, e.g. one-, two-, and infinity dimensions, in Minkowski space derives less clusters, while the assignment of a larger dimension results in a larger cluster number with a more detailed segmentation.

[1]  T. M. Lillesand,et al.  Remote Sensing and Image Interpretation , 1980 .

[2]  P. N. Suganthan,et al.  Robust growing neural gas algorithm with application in cluster analysis , 2004, Neural Networks.

[3]  Fawwaz T. Ulaby,et al.  An automated unsupervised/supervised classification methodology , 1998, IGARSS '98. Sensing and Managing the Environment. 1998 IEEE International Geoscience and Remote Sensing. Symposium Proceedings. (Cat. No.98CH36174).

[4]  Ashraf A. Kassim,et al.  Character recognition using statistical moments , 1999, Image Vis. Comput..

[5]  Ujjwal Maulik,et al.  Performance Evaluation of Some Clustering Algorithms and Validity Indices , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  G. Borgefors,et al.  Integrated method for boundary delineation of agricultural fields in multispectral satellite images , 2000, IGARSS 2000. IEEE 2000 International Geoscience and Remote Sensing Symposium. Taking the Pulse of the Planet: The Role of Remote Sensing in Managing the Environment. Proceedings (Cat. No.00CH37120).

[7]  Krištof Oštir,et al.  Application of satellite remote sensing in natural hazard management: The Mount Mangart landslide case study , 2003 .

[8]  Sushmita Mitra.  An evolutionary rough partitive clustering , 2004, Pattern Recognit. Lett..

[9]  S. Hsu,et al.  Application of remotely sensed data to the assessment of terrain factors affecting the Tsao-Ling landslide , 2004 .

[10]  T. M. Lillesand,et al.  Remote sensing and image interpretation. Second edition , 1987 .

[11]  Du-Yih Tsai,et al.  Medical image classification using genetic-algorithm based fuzzy-logic approach , 2004, J. Electronic Imaging.

[12]  T. Ross.  Fuzzy Logic with Engineering Applications , 1994 .

[13]  Dirk Roose,et al.  MIMD Divide-and-Conquer Algorithms for the Distance Transformation. Part II: Chamfer 3-4 Distance , 1995, Parallel Comput..

[14]  Horst Martini,et al.  Median and center hyperplanes in Minkowski spaces--a unified approach , 2001, Discret. Math..

[15]  Lin-Yu Tseng,et al.  A genetic approach to the automatic clustering problem , 2001, Pattern Recognit..

[16]  Kevin J. Parker,et al.  Clustering approach to bone and soft tissue segmentation of digital radiographic images of extremities , 2003, J. Electronic Imaging.

[17]  Azriel Rosenfeld,et al.  Centers of sets of pixels , 2000, Discret. Appl. Math..

[18]  Dirk Roose,et al.  MIMD Divide-and-Conquer Algorithms for the Distance Transformation. Part I: City Block Distance , 1995, Parallel Comput..

[19]  K. Swanepoel.  Cardinalities of k -distance sets in Minkowski spaces , 1999 .

[20]  C. A. Murthy,et al.  In search of optimal clusters using genetic algorithms , 1996, Pattern Recognit. Lett..

[21]  Ming-Der Yang,et al.  A genetic algorithm (GA) based automated classifier for remote sensing imagery , 2007 .

[22]  Fernando Moura-Pires,et al.  Modeling proportional membership in fuzzy clustering , 2003, IEEE Trans. Fuzzy Syst..

[23]  G. Beni,et al.  A new fuzzy clustering validity criterion and its application to color image segmentation , 1991, Proceedings of the 1991 IEEE International Symposium on Intelligent Control.