作者: Tobias Klinder , Samuel Kadoury , Cristian Lorenz
DOI: 10.1007/978-3-319-03813-1_11
关键词:
摘要: Modeling of the human spine requires extension from single object modeling to ensembles. The consists a constellation vertebrae where individual show complex shape. While most neighbouring look very similar, their shape changes significantly along spine. Due these challenges, more sophisticated model formulations are needed that go beyond vertebrae. In this article, we combine several high-level models into one common framework. represented as set covering shape, gradient and appearance information well relative location orientation. By encoding further anatomical representation vertebrae, e.g., important regions or significant landmarks, clinically relevant parameters can be easily derived models. is expressed sequence rigid transformations between different statistical methods used cover variability spinal curvatures. For selected applications vertebra labelling in limited field view scans segmentation both CT MRI, how comprehensive framework for an automatic image interpretation medical images Furthermore, problem change assessment osteoporotic fracture detection tackled with example CAD.