Outlier Rejection for Adaptive Neonatal Segmentation

Keywords

abnormalities accuracy adaptive adults amount analyzed anatomical anatomy artifacts assess assumes assuming atlas atlases attempt babies basis belongs bias birth blue brain cerebellum certain characteristic checks chosen class clinical clusters cohort combination combined component conjugate constraining containing contrast control corrected correction correlation cortical cross deep dependence depends described developments dice digital distance distribution done enables example excessive existence extraction field fluid frequency functions fundamental furthermore fusion gestation gestational green hyper hypo identify ignores improvement improvements includes incorporating influence injury intensities intensity introduced kingdom label lack lead likelihood linear manual manually markedly matrix maximization mitigate model modeled much multivariate nature neighbors neonatal normalized novel outliers pathological pons population presence prior priori priors probabilistic probability problem proposed providing quality random recent reducing regards rejection respectively reversal risk sample score segmentation skull smoothness space spatial staple stripped stripping strong structures studies subset substantial term third tissue tissues uniformity variability varying ventricles volumetric weeks white