Published Paper Details:

Breast cancer is the leading cause of cancer death among women. Screening x-ray mammography is the first and regular method currently available for the reliable early detection and treatment can be started for hopeful cure of breast cancer. Research indicates that the mortality rate could decrease by 35% if women age around 45-50years and older have regular mammograms. The detection rate can be confirmed 15-30% by providing the radiologist with results from a computer-aided diagnosis (CAD) system acting as a second opinion. However, among screening mammograms routinely interpreted by radiologists, very few (approximately 10%) cases actually have breast cancer. It would be beneficial if an accurate CAD system existed to identify normal mammograms and thus allowing the radiologist to focus on suspicious cases. This strategy could reduce the radiologist's workload and improve screening performance. Since mammography is considered as the most effective means for breast cancer diagnosis, this paper introduces two separate techniques for mass and micro-calcification segmentation in digital mammograms. There are two different segmentation techniques used for masses and microcalcifications.For masses there are three steps- background subtraction, fuzzy texture representation and entropic theresholding. Similarly micro-calcifications are also segmented in three stages - background subtraction, Laplacian of Gaussian filtering and contrast estimation followed by thresholding. GLCM texture feature extractions are known to be the most common and powerful techniques for texture analysis including histogram statistical features. Experiments have been taken for a data set of 322 images taken from MIAS of different types with the aim of improving the accuracy of Association Rule classifier by generating minimum no. of rules to cover more patterns. In this study, an incremental updating technique is applied to associative classification for constructing classification system when a new training dataset is appended to an old training dataset. The proposed algorithm, called Incremental Classification Based on Association Rules (ICBA). ICBA has 2 phases which are rule generator phase (ICBA-RG) and classifier building phase (ICBA-CB). In order to reduce the execution time, we applied the concept of Fast Update algorithm (FUP) algorithm to both phases of our algorithm. The experiment results show that the proposed algorithm has execution time better than CBA algorithm.