FUZZY RELATIONAL PATTERN DYNAMICS OF THE BREAST CANCER CELLS Z. Zeybek 1 , Z. Telatar 2 , Y. Öztürk 2 Ankara University, Engineering Faculty Dept. of Chemical 1 and Electronics 2 Engineering E-mail: telatar@science.ankara.edu.tr Abstract: In this study, the microcalcification modeling problem has been handled to detect the microcalcification clusters. The detection of microcalcifications is carried out over the filtered mammogram image. It is observed that, the histograms of the band-pass filtered subimage are very close to the Gaussian distribution. Our detection scheme based on these parameters are modeled by fuzzy relational matrix estimating microcalcifications automatically in each square region. If a region has high positive skewness and kurtosis then it is marked as a region of interest . Both the skewness and the kurtosis assume very small values in the healty breast region, while they have high values in the region containing microcalcifications. Experimental results confirm our detection scheme. Keywords: Mammography, microcalcifications, HOS parameters, pattern recognition, fuzzy logic 1. Introduction: Generally, the meaning of the word “cancer” is introduced in literature as increasing of the cells without control. In regard of the working of the cells differ from the normal cells in human body. Sometimes, the cells do not work properly and they also begin to do some new duties. These cells begin to increase anormally and prevent from their duties in the region invaded. The forming period of the cancer cells can show a changeability according to the sort of cancer. The type of cancer is called by means of an organ and its tissue on which the cancer cells take on roots. Breast cancer is one of the most deadly disease for middle-aged woman. One out of eight woman is prone to this disease in her life time. There are many risk factors to to be caught of the breast cancer. The most important way to prevent from the breast cancer is to supply early detection. It extracts the cancer cells which are hidden clinically and formed in early phase, instead of the lesions determined by handle. Microcalcifications appearing as small, bright spots on mammograms are tiny calcium deposits in breast parenchymal tissue structures. Since the microcalcifications in mammograms are small and subtle abnormalities, they may be overlooked by a radiologist with a magnifying glass to find out tumors such as microcalcifications, masses and stellate regions [1].. In mammography, mass, stellate lesions and microcalcification clusters are detected and classified by using feature extraction, feature selection, and classification steps. They can be processed by using the Fourier transform, the wavelet transform or the non-linear filter based transforms, higher order correlations, autocorrelations. Energy and entrophy computations of the signal are some of