AN ENHANCED DIAGNOSIS METHOD FOR CANCER TREATMENT USING FUZZY MODEL

Computer aided diagnosis (CAD) systems evaluate the conspicuous structure. Most of the existing diagnosis methods used mammogram mass characteristics but in the proposed technique microcalcification (MC) characteristics are used. In this project, cancer cells are diagnosis using the steps like mammogram image, pre-processing techniques, segmentation, classification, fuzzy logic and Receiver Operating Characteristics (ROC) curve analysis. Mammogram images are collected from the National Institute of Cancer. Pre-processing technique performed using normalization and median filtering. Based on the local threshold method, segmentation is done to identify the location of microcalcification present cells. Classification technique includes Perceptron algorithm and case based reasoning (CBR) method. Perceptron algorithm is used to classify the microcalcification present cells and microcalcification absent cells.CBR is used to classify the microcalcification present cells into following classes initial, very small, small, medium, high, very high. Fuzzy logic used for decision making purpose. And system performance analyzed using ROC curve analysis. This method provides better performance (95%) compare to the previous diagnosis techniques.