We present a registration-based semiautomatic mandible segmentation (SAMS) pipeline designed to process a large number of computed tomography studies to segment 3-dimensional mandibles.
The pipeline consists of a manual preprocessing step, an automatic segmentation step, and a final manual postprocessing step. The automatic portion uses a nonlinear diffeomorphic method to register each preprocessed input computed tomography test scan on 54 reference templates, ranging in age from birth to 19 years. This creates 54 segmentations, which are then combined into a single composite mandible.
This pipeline was assessed using 20 mandibles from computed tomography studies with ages 1 to 19 years, segmented using both SAMS-processing and manual segmentation. Comparisons between the SAMS-processed and manually-segmented mandibles revealed 97% similarity agreement with comparable volumes. The resulting 3-dimensional mandibles were further enhanced with manual postprocessing in specific regions.
Findings are indicative of a robust pipeline that reduces manual segmentation time by 75% and increases the feasibility of large-scale mandibular growth studies.
From the *Vocal Tract Development Lab, Waisman Center, and †Department of Biostatistics and Medical Informatics, University of Wisconsin-Madison, Madison, WI.
Received for publication March 30, 2017; accepted July 12, 2017.
Correspondence to: Houri K. Vorperian, PhD, Vocal Tract Development Lab, Waisman Center, University of Wisconsin-Madison, 1500 Highland Ave, Room 427, Madison, WI 53705 (e-mail: email@example.com).
This work was supported by NIH research grant R01 DC006282 (MRI and CT Studies of the Developing Vocal Tract, Houri K. Vorperian, Principal Investigator) from the National Institute on Deafness and other Communication Disorders, and by a core grant P30 HD03352 and U54 HD090256 to the Waisman Center (Albee Messing, principal investigator) for research support from the National Institute of Child Health and Human Development.
All authors declare no conflict of interest.