Related Experiment Video
Updated: Jul 5, 2025
09:10
Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
Published on: August 5, 2021
1.8K
Accurate malocclusion tooth segmentation method based on a level set with adaptive edge feature enhancement.
Shuyi Jiang1, Han Zhang2, Zhi Mao2
1College of Computer Science and Technology, Changchun University of Science and Technology, Changchun, 130012, China.
Heliyon
|January 23, 2024
Summary
This study introduces a novel method for precise tooth segmentation in complex oral conditions, achieving high accuracy even with dental crowding and malocclusion. The approach offers robust segmentation for improved dental research and clinical practice.
Area of Science:
- Oral and Maxillofacial Imaging
- Medical Image Analysis
- Computational Anatomy
Background:
- Accurate tooth segmentation is crucial for dental diagnostics and treatment planning.
- Complex oral conditions like malocclusion and crowding pose significant challenges for automated segmentation methods.
- Existing segmentation techniques often struggle with intricate anatomical variations and close tooth proximity.
Purpose of the Study:
- To develop and validate a robust three-step method for accurate tooth segmentation in Cone-Beam Computed Tomography (CBCT) images.
- To address challenges posed by complex structural interference, tooth rotation, and displacement in dental crowding.
- To provide a reliable tool for segmenting individual teeth under difficult oral conditions.
Main Methods:
- A three-step segmentation approach utilizing global and local level-set models.
- Extraction of bony tissue using a global convex level-set model.
- A flexible curve extraction method for separating adjacent teeth with structural interference.
- A local level-set model with adaptive edge feature enhancement for precise individual tooth segmentation, particularly for rotated or displaced teeth.
Main Results:
- High segmentation accuracy demonstrated across different tooth types: incisors (93.30%), canines (93.47%), premolars (93.24%), and molars (93.89%) using Dice similarity coefficient.
- Low average tooth centroid distances, indicating precise localization: 0.66 mm (incisors), 0.61 mm (canines), 0.87 mm (premolars), and 0.80 mm (molars).
- The method proved effective without requiring highly precise annotated datasets.
Conclusions:
- The proposed method effectively segments teeth in complex oral scenarios, including those with malocclusion and crowding.
- The approach exhibits superior robustness compared to other methods, handling structural interference and tooth displacement.
- The findings offer valuable data for advancing dental research and clinical applications.
