Related Experiment Video
Updated: Jul 12, 2025
Author Spotlight: Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans
Published on: September 8, 2023
[Deep learning-assisted construction of three-dimensional face midsagittal plane based on point clouds].
1Center of Digital Dentistry, Faculty of Prosthodontics, Peking University School and Hospital of Stomatology & National Center for Stomatology & National Clinical Research Center for Oral Diseases & National Engineering Research Center of Oral Biomaterials and Digital Medical Devices & Research Center of Engineering and Technology for Computerized Dentistry Ministry of Health & Beijing Key Laboratory of Digital Stomatology, Beijing 100081, China.
A novel Dynamic Graph Registration Network (DGRNet) algorithm automates 3D facial midsagittal plane construction. This intelligent registration method offers improved efficiency and accuracy for clinical digital design and analysis.
Area of Science:
- Medical Imaging
- Computer Vision
- Biomedical Engineering
Context:
- Accurate 3D facial midsagittal plane construction is crucial for clinical digital design and analysis.
- Existing methods may lack efficiency and precision in automated registration.
- The development of intelligent algorithms is needed to address these challenges.
Purpose:
- To establish an intelligent registration algorithm for automatic 3D facial midsagittal plane construction.
- To develop and evaluate the Dynamic Graph Registration Network (DGRNet) for original-mirror alignment.
- To provide a reliable reference for clinical applications.
Summary:
- The Dynamic Graph Registration Network (DGRNet) was developed using point cloud data from 200 patients.
- The algorithm utilizes feature vectors, singular value decomposition, and principal component analysis for registration.
- Evaluation on 50 clinical cases showed an average angle error of 1.05°±0.56°, with a 90% success rate within 2.00°.
Impact:
- The DGRNet offers a new, intelligent solution for 3D facial midsagittal plane construction.
- This method enhances the efficiency and effectiveness of craniofacial digital design and treatment planning.
- The algorithm provides a valuable tool for clinicians in digital dentistry and orthodontics.
