Inferior alveolar canal segmentation based on cone-beam computed tomography

Area of Science:

• Dentistry
• Medical Imaging
• Computer Vision

Background:

• The inferior alveolar canal (IAC) is crucial for dental procedures.
• Iatrogenic injury risk necessitates precise IAC analysis.
• Cone-beam computed tomography (CBCT) provides detailed anatomical data.

Purpose of the Study:

• To develop and validate an automated segmentation method for the inferior alveolar canal (IAC) in CBCT images.
• To enhance the accuracy of IAC shape and position analysis for dental clinicians.
• To reduce the risk of iatrogenic injury during dental procedures.

Main Methods:

• Preprocessing of CBCT images including Hounsfield unit clipping and gray normalization.
• Generation of curved multi-planar reconstruction (MPR) image sets using a smoothed dental arch curve.
• K-means clustering of texture parameters from a gradient-enhanced gray level-gradient co-occurrence matrix.
• Segmentation of IAC edges using 2D line-tracking and smoothing with a fourth-order polynomial.

Main Results:

• The proposed method achieved high segmentation accuracy on 21 clinical CBCT datasets.
• Quantitative evaluation showed Dice Similarity Index (DSI) of 0.93 for both left and right IACs.
• Average Symmetric Surface Distance (ASSD) and Mean Curve Distance (MCD) were consistently low, indicating precise segmentation.

Conclusions:

• The developed method offers an effective solution for IAC image enhancement and segmentation.
• Accurate analysis of IAC morphology and spatial relationships is achievable without superimposition or distortion.
• The method successfully segments smooth IAC edges, aiding in safer dental interventions.