Anatomical prior-based vertebral landmark detection for spinal disorder diagnosis

Yukang Yang, Yu Wang, Tianyu Liu, Miao Wang, Ming Sun, Shiji Song, Wenhui Fan, Gao Huang

Abstract: Highlights•We integrate priors to detect accurate and anatomically plausible vertebral landmarks.•We introduce an anatomy-constraint loss to regularize training process with priors.•We propose a simple-yet-effective anatomy-aided inference procedure.•We provide novel numerical metrics for assessing anatomical plausibility.•We employ our prior-based method for scoliosis assessment and fracture classification.