bilinear PCA

This work introduces a data-driven model of pediatric cranial bone development during the first two years of life. We present an automatic algorithmic pipeline to segment the cranial bones from a large retrospective cross-sectional dataset of computed tomography images of normative pediatric subjects, establish local anatomical correspondences between crania guided by the cranial sutures, and create a statistical model of the anatomical variability of the calvaria and its normal temporal changes during development. Our data-driven statistical approach assumes temporal continuity of cranial development to avoid assumptions about the biophysical processes involved with bone growth. read more