Segmentation of the Oral and Facial Regions from Imaging Modalities with Reduced or No Ionizing Radiation

Abstract

Reducing diagnostic radiation has become a major concern in clinical orthodontic treatment. We present two approaches for segmenting the human mandibular body from magnetic resonance (MR) images and segmenting the human anterior teeth from cone beam computer tomography (CBCT) images. Both of the segmentation approaches allow clinicians to study the oral and maxillofacial images in modalities with relative lower radiations. A two-stage rule-constrained seedless region growing approach was presented to segment the mandibular body in MR images. The proposed approach was implemented and the segmentation results were compared with other algorithms and the ground truth. The proposed method showed the best results in most scenarios. A level-set based approach for segmenting the anterior tooth segmentation from CBCT was then presented. The proposed approach was implemented and the segmentation results were compared with other algorithms and the ground truth. The proposed method showed the best results in most scenarios.