Related Experiment Video
Updated: Feb 22, 2026
09:10
Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
Published on: August 5, 2021
2.3K
Assessment of automatic segmentation of teeth using a watershed-based method.
Antoine Galibourg1,2, Jean Dumoncel1, Norbert Telmon1,3
11 Laboratoire Anthropologie Moléculaire et Imagerie de Synthèse, Université Paul Sabatier, Toulouse, France.
Dento Maxillo Facial Radiology
|September 23, 2017
Summary
Automatic segmentation (AS) of teeth using a watershed method shows reliable volumetric measurements comparable to semi-automatic segmentation (SAS) with micro-CT data. While CBCT data shows increased discrepancies with voxel size, AS remains valuable for clinical dentistry with minor manual adjustments.
Area of Science:
- Dental Imaging and Diagnostics
- Computational Anatomy
- Medical Image Analysis
Background:
- Automatic segmentation (AS) of teeth is crucial for research and clinical applications.
- Validated semi-automatic segmentation (SAS) methods exist, but AS offers potential time savings.
- Evaluating the accuracy and reproducibility of AS in 3D reconstructions is essential.
Purpose of the Study:
- To assess the impact of watershed-based automatic segmentation (AS) on 3D tooth reconstruction accuracy and reproducibility.
- To compare AS with validated semi-automatic segmentation (SAS) for volumetric measurements.
- To analyze the influence of different imaging modalities (micro-CT and CBCT) and voxel sizes on segmentation performance.
Main Methods:
- 52 teeth were scanned using micro-CT (41 µm) and CBCT (76, 200, 300 µm).
- Teeth were segmented using both watershed-based AS and SAS.
- Volumetric measurements and statistical analyses were performed; surfaces were aligned using SAS as reference, and discrepancies were visualized using color maps.
Main Results:
- AS reconstructions yielded comparable tooth volumes to SAS for high-resolution micro-CT data (41 µm).
- Volume differences increased with larger voxel sizes in CBCT data.
- Maximum discrepancies were observed at cervical margins and incisal edges, though the overall tooth form was preserved.
Conclusions:
- Watershed-based AS is time-efficient for micro-CT data in dental research, providing accurate segmentation.
- AS with CBCT data can represent the general tooth form but requires manual refinement for metrically reliable clinical measurements.
- The AS method shows potential for clinical dentistry applications when combined with targeted manual adjustments.
