Semiautomated Three-Dimensional Craniofacial Imaging and Modeling
RICHARD A. LEVY · Investigative Radiology · 1994
RATIONALE AND OBJECTIVES: The author identified clinical parameters and developed thresholding strategies that facilitate semiautomated three-dimensional craniofacial imaging and modeling. METHODS: A polyvinylchloride phantom was imaged on a computed tomography (CT) scanner using multiple scan planes, surrounding media, and two-dimensional reconstruction filters. Three-dimensional reconstructions were performed by an experienced operator on a commercial workstation. A simple mathematical thresholding algorithm derived from these trials was applied to 17 embryologic craniofacial specimens and compared with operator-selected thresholds in the generation of three-dimensional CT images. RESULTS: Utilization of the parameter alpha provides an experimental framework in which operator-selected thresholds can be used to refine simple mathematical thresholding formulae. We demonstrated significant changes in three-dimensional image accuracy based on CT scan plane. CONCLUSIONS: Experimentation with a CT phantom enables semiautomated three-dimensional craniofacial imaging and modeling. Further investigation of two-dimensional CT craniofacial datasets is justified.