Modeling of the human upper airway from multimodal 3D dentofacial images

Abstract

A patient-specific virtual upper airway model is important for clinical, education, and research applications. Cone beam computed tomography (CBCT) and magnetic resonance (MR) are popular imaging modalities of the upper airway but automatic segmentation of these images is still challenging due to noise and the complex anatomical features, especially at the region of the nasal cavity and paranasal sinuses. In this thesis, we describe segmentation schemes for the segmentation of nasal cavity and paranasal sinuses from CBCT images and the nasal passage from MR images. First, we present a multi-step level set segmentation scheme based on Kullback-Leibler divergence and anisotropic localized active contour for the automatic segmentation of CBCT images to create a 3D model of the nasal cavity and paranasal sinuses. Second, we propose a graph-based method for the segmentation of thin volumetric structures and apply it to segment the nasal passage from head MR images. Third, we register the 3D surface mesh of the dental study model to the head MR images to create the entire upper airway model using only non-radiation imaging modalities.