A Bayesian Network Based Classification of Breast Lesion in Digital Mammogram

Breast cancer is a serious issue in the worldwide females. Breast cancer is the type of cancer which develops from breast cells. For detection of breast cancer there are different types of screening techniques are available. For detection of breast cancer this paper includes several techniqus. In this paper, the first step was to remove noise from the image; median filter is used to remove the unwanted noise in the image. In the MIAS database pectoral muscles are available in the image, pectoral muscle are removed by calculating the thresholding value of an image.The entropy based segmentation approach is proposed to segment a gray-scale breast image. The approach calculates the histogram of an image also finds the entropy value of image. Then by finding the thresholding value of an image the segmented image is shown at the output. In this paper, an efficient and fast entropic method for noisy cell image segmentation is presented. Then the features like mean, standard deviation, Entropy, Skewness, Kurtosis, Variance, Energy, Correlation, Smoothness and Root mean square are extracted from a segmented image, this features are then given to the input of Bayesian network to classify the image according to the feature value. Experimental results show that the proposed method is efficient and much more tolerant to noise than other techniques.