Segmentation of the Breasts in MR-Mammography

Introduction Contrast enhanced MR mammography is a highly sensitive method in breast cancer detection. Tracking the inflow and washout of contrast agents is used to classify suspect tissue as malign or benign. Segmenting the breasts is needed in order to: • quantify the tissue in fat and parenchyma, • make local elastic motion correction [1] faster and more robust by restricting the registration to the tissue of interest, • prepare the data for improved volumetric display such as volume rendering. Standard segmentation strategies like region growing and closed snakes cannot handle this appropriately. Thus an algorithm is proposed which utilizes image features of MR mammograms efficiently.