Related Experiment Video
Updated: Sep 10, 2025
Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
Published on: August 5, 2021
Automatic margin line extraction using 3D deep learning on digital surface models of prepared teeth for crown
Ammar Alsheghri1, Ying Zhang2, Julia Keren3
1Mechanical Engineering Department, King Fahd University of Petroleum and Minerals (KFUPM), Dhahran, 31261, Kingdom of Saudi Arabia; Interdisciplinary Research Center for Biosystems and Machines, King Fahd University of Petroleum and Minerals (KFUPM), Dhahran, 31261, Kingdom of Saudi Arabia.
This study introduces an AI-powered digital solution for automatic margin line creation in dental crown design, improving consistency and efficiency over manual methods. The artificial intelligence framework accurately extracts margin lines from 3D scans of prepared teeth.
Area of Science:
- Digital Dentistry
- Artificial Intelligence in Healthcare
- Biomedical Engineering
Background:
- Manual margin line design in dental crowns is subjective and lacks repeatability.
- Current digital methods for margin line creation can be time-consuming and inconsistent.
- Accurate margin line definition is crucial for successful dental restorations.
Purpose of the Study:
- To develop and validate a personalized digital solution for automatic and consistent margin line creation on prepared teeth.
- To leverage supervised deep learning for accurate segmentation of prepared teeth in 3D scans.
- To establish an end-to-end artificial intelligence framework for margin line extraction in dental crown design.
Main Methods:
- Utilized supervised deep learning on a dataset of 1113 3D scanned prepared teeth.
- Employed a majority voting classifier with 5-fold cross-validation to enhance segmentation performance.
- Implemented a novel post-processing procedure for margin line correction and extraction, followed by comparison with ground truth margins.
Main Results:
- The AI model successfully extracted margin lines in 70 out of 78 test cases within a 200 μm threshold and 62 within a 100 μm threshold.
- The proposed post-processing procedure demonstrated superior accuracy (DSC = 0.986) compared to the traditional graph-cut technique (DSC = 0.974, p=0.024).
- Achieved high accuracy and consistency in margin line segmentation, outperforming existing methods.
Conclusions:
- The proposed end-to-end AI framework offers a consistent and time-efficient digital tool for automatic margin line extraction.
- This artificial intelligence solution has the potential to significantly improve the workflow for dental crown design.
- The technology provides a repeatable and accurate method for defining critical margins in digital dentistry.
