Segmentation and Classification of Masses in Digital Mammograms for Diagnosing Breast Cancer

Early diagnosis of breast cancer helps to improve the successful treatment.Mammographic images are obtained from X-ray instrumentation. They are noisy and poor contrast due to ill performance of the hardware systems. Enhancement helps in improving the contrast of mammograms and reduces the false positive findings. Various enhancement techniques are available in image processing for improving the contrast of the mammograms. The Polynomial Volterra filtering technique is applied for the enhancement. This technique is the combination of Type-0 and Type-II polynomial filters which gives better enhancement results on mammograms with different background tissues. The genetic algorithm is applied for the segmentation of tumor region from mammograms. The segmented results contain pectoral muscles and the tumor region. The pectoral muscles are removed by applying the region properties of the image. The tumor region is obtained at the segmented result.Breast cancer classification system is developed for classifying the lesions into benign and malignant. Features concerning the shape and the texture of the lesion regions are presented to SVM and Adaboost classifiers for classifying the tumor effectively. Keywords—Breastcancer detection, contrast enhancement, genetic algorithm, polynomial filtering.