Automated tracking of the mandibular canal in CBCT images using matching and multiple hypotheses methods

Abstract

A method for the automated tracking of the mandibular canal in Cone Beam Computed Tomography images is presented. In contrast to most proposed methods a fully automated method is presented including the detection of the mental and mandibular foramen. Extraction of these points is done by searching for high-contrast gaps in the mandible. The path of the mandibular canal can then be tracked in between these two points using a matching technique consisting of a template function which maps spatial coordinates to a cost. The robustness of the tracking is increased by recursively tracking multiple hypotheses.