A NOVEL APPROACH FOR DEVELOPING A COMPLETE AUTOMATIC 3D CEPHALOMETRIC ANALYSIS AND AUTOMATIC MANDIBLE, MAXILLA AND DENTITION SEPARATION SYSTEM

Abstract

Traditional cephalometric analysis is performed by tracing radiography landmarks on Lateral X-ray

images. Generally, the orthodontist uses these landmarks to measure the linear and angular values needed for both of the pretreatment planning and the treatment evaluation. Traditional cephalometric is time consuming and error prone. The 2D cephalometric requires symmetry of anatomic landmarks, and it is less useful in evaluating patients with severe craniofacial anomalies. In addition, the orthodontists solve 3D problem by using only 2D images, which is not sufficient. Considering, the recent advent of computer-assisted tomography (CT) and the ability to reconstruct 30 model for the anatomical parts, a good refinement in preoperative planning for many surgical procedures and treatment planning had been achieved.

The major contribution of this work lies in building Digital Reconstructed Radiography (ORR) images from patient's CT scans. Therefore, two image could be built and reconstructed that represent the orthogonal X-ray, the lateral and frontal cephalogram. Furthermore, the Active Appearance Models (AAM) are applied on both the lateral and frontal X-ray images, which result in a fully automatic annotation for the 30 cephalometric landmarks on the head model. This work can also be used to segment the lower and upper parts of patient's jaw, both the mandible and maxilla respectively. The Mandible and Ma illa separation is the first step towards their manipulation surgically or orthodontically. This approach can be used on symmetric or asymmetric patient's skull. The automatic teeth separation is performed by using Combined Threshold Technique-Connected Component Algorithm. This work results in a system that performs an automatic analysis for the 3D patient skull. It can be used to automate pre-treatment planning process by applying the automatic extraction and visualization for the jaw parts and teeth with accurate and reliable analysis.