Automatic Segmentation of Digital Orthopantomograms for Forensic Human Identification

Dental radiographic images are one of the most popular biometrics used in the process of forensic human identification. This led to the creation of the Automatic Dental Identification System with the goal of decreasing the time it takes to perform a single search in a large database of dental records. A fully automated system identifying people based on dental X-ray images requires a prior segmentation of the radiogram into sections containing a single tooth. In this paper, a novel method for such segmentation is presented, developed for the dental radiographic images depicting the full dentition — pantomograms. The described method utilizes the locations of areas between necks of teeth in order to determine the separating lines and does not depend on the articulation of gaps between adjacent teeth, thus improving the results achieved in the situation of severe occlusions.