Classification of Microcalcifications Using Multi-Dimensional Genetic Association Rule Miner

Breast cancer is a most common disease diagnosed in women. The Microcalcification Clusters (MCs) in the mammograms are one of the important early sign. The accurate detection of microcalcifications is a key problem in Computer Aided Detection (CAD). In this paper, we have proposed a novel association rule mining approach for classification of microcalcifications. Initially, the shape features are extracted from the digital mammograms. With these feature values association rules are constructed to develop a rule based system for classification of microcalcifications. A novel Multidimensional Genetic Association Rule Miner (MGARM) is proposed for rule construction. The result shows that the proposed rule-based approach reaches the classification accuracy over 85% and also demonstrates the use and effectiveness of association rule mining in image classification. Index terms – Mammograms, Microcalcification Clusters, Shape Features, Classification, Association Rule Mining, Genetic Algorithm.