Jaw tissues segmentation in dental 3D CT images using fuzzy-connectedness and morphological processing

Area of Science:

• Oral and Maxillofacial Surgery
• Medical Imaging
• Computational Anatomy

Background:

• Accurate pre-operative planning is crucial for successful oral surgery and dental implant placement.
• Precise segmentation of jaw anatomy, including teeth, cortical bone, trabecular bone, and the inferior alveolar nerve, is essential for this planning.

Purpose of the Study:

• To present a novel, automated method for segmenting jaw tissues from computed tomography (CT) data.
• To accurately delineate the inferior alveolar nerve path and surrounding jaw structures for surgical planning.

Main Methods:

• The method employs fuzzy connectedness object extraction and mathematical morphology processing.
• It utilizes CT data to generate pseudo-orthopantomographic views for nerve path estimation and cross-sectional views for tissue segmentation.

Main Results:

• The automated method achieved promising results on a dataset of over 9000 cross-sections from 20 patients.
• Evaluation using Jaccard index (0.726±0.031) and Dice's coefficient (0.840±0.019) demonstrated high segmentation accuracy.
• Point-to-point (0.144±0.023 mm) and point-to-curve (0.163±0.025 mm) distances confirmed the method's precision.
• The system demonstrated automation with approximately 5% error relative to the nerve's diameter.

Conclusions:

• The developed automated method provides accurate and efficient segmentation of jaw tissues and the inferior alveolar nerve.
• Its high accuracy and automation make it easily integrable into existing dental planning systems, enhancing pre-surgical preparation.