Related Experiment Video
Updated: Jun 21, 2025
09:10
Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
Published on: August 5, 2021
1.7K
Convolutional neural network for automated tooth segmentation on intraoral scans
Xiaotong Wang1,2, Khalid Ayidh Alqahtani3, Tom Van den Bogaert1
1OMFS IMPATH Research Group, Department of Imaging and Pathology, Faculty of Medicine, KU Leuven, Kapucijnenvoer 33, Leuven, 3000, Belgium.
BMC Oral Health
|July 16, 2024
Summary
This study introduces a convolutional neural network (CNN) model for accurate automatic tooth segmentation on intraoral scans. The developed 3D U-Net pipeline significantly improves efficiency and reliability in digital dentistry workflows.
Area of Science:
- Computer Vision
- Medical Imaging
- Digital Dentistry
Background:
- Automatic tooth segmentation from intraoral scans (IOS) is crucial for digital dentistry.
- Existing methods struggle with the variability of dental conditions.
- A robust automated solution is needed for clinical applications.
Purpose of the Study:
- To develop and evaluate a convolutional neural network (CNN) model for automatic tooth segmentation on IOS images.
- To assess the performance and clinical applicability of the proposed model.
Main Methods:
- A dataset of 761 IOS images was used.
- A multi-step 3D U-Net pipeline was designed for automated segmentation.
- Performance was evaluated by accuracy and time, with clinical applicability tested on a cloud platform.
Main Results:
- The CNN model achieved a 91% Intersection over Union (IoU) score.
- Automated segmentation averaged 31.7 seconds per jaw.
- A refined AI approach (R-AI) significantly reduced segmentation time and improved reliability compared to semi-automatic methods.
Conclusions:
- The 3D U-Net pipeline provides accurate, efficient, and consistent automatic tooth segmentation for IOS data.
- The cloud-based platform offers a viable solution for IOS segmentation in clinical settings.
