Cluster Approach to Pattern Recognition
This paper deals with the mathematical morphology as a cluster approach to image processing and the Hough transformation as an object of statistical physics. It is shown that the operation in mathematical morphology is equivalent to a single step of dynamics in a special neural network. We investigate dynamics of an image (the neural network) and examine the relation between the time interval to reach to an associative memory and the threshold. As for the Hough transformation, we derive the free energy expression for the transformation and proposed the Gaussian sum method.
KeywordsPartition Function Cluster Approach Associative Memory Image Restoration Mathematical Morphology
Unable to display preview. Download preview PDF.
- 3.See for example, Foundations and Applications of Cluster Variation Method and Path Probability Method, edited by T. Morita, M. Suzuki, K. Wada, and M. Kaburagi, Progr. Theor. Phys. Suppl. 113 (1994).Google Scholar
- 4.K. Fukunaga, Introduction to Statistical Pttern Recognition, (Academic Press, New York, 1972).Google Scholar
- 5.T. Geszti, Physical Models of Neural Network, (World Scientfìc, Singapore, 1990).Google Scholar
- 10.K. Tanaka, in Theory and Application of the Cluster Variation and Path Probability Methods, edited by J. L. Morán-López and J. M. Sanchez (Plenum, New York, 1996), p. 345.Google Scholar
- 11.J. Serra, Image Analysis and Mathematical Morphology, Vol. 1, (Academic Press, New York, 1989).Google Scholar
- 14.P. V. C. Hough, U. S. Patent 3069654 (1962).Google Scholar
- 17.Q. Ou and A. Ono, preprint.Google Scholar