Automatic Template-Based Breast Segmentation on MRI Using Nonrigid Registration Algorithms

Keywords

able accuracy adjacent aided analyzing applied automatic background bilateral body boundary bounding breast breasts cancer central chemotherapy chest china closely clustering cohort computer connected consistently constrains consuming continues contour correction correlation cover cropped curve deformation deformed demons density described detected detection determine develop developed diagnosis directly either ensure example exclude excluded extracted features field fifthly final find firstly fitting five fuzzy generation green hand heart help highlighted hospital identified inferiorly inhomogeneity initial input inside intensity kernel landmarks lastly lateral lesion locate location lung manual manually margin match medical model moderate muscle muscles mussel necessary nonrigid often outer outline patterns performance phys planned posterior preserve previously process processing produce quantitative radiology registered registration remove risk robustness scheme segment segmentation segmented selected semi served shape shapes sides similarity since sixthly slice slices smooth smoothed spine step subject subtracting successfully suitable superiorly supported template thirdly thoracic tissue tissues tumor usually variety various wall yellow