Related Experiment Video
Updated: Jun 11, 2025
10:23
Author Spotlight: Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans
Published on: September 8, 2023
2.6K
Clinical validity and precision of deep learning-based cone-beam computed tomography automatic landmarking algorithm
Jungeun Park1, Seongwon Yoon2,3, Hannah Kim3,4
1Department of Orthodontics, College of Dentistry, Yonsei University, Seoul, Korea.
Imaging Science in Dentistry
|October 7, 2024
Summary
A new deep learning algorithm for automatic landmarking in cone-beam computed tomography (CBCT) shows accuracy comparable to manual methods. This artificial intelligence approach significantly reduces landmark identification time, improving diagnostic efficiency.
Area of Science:
- Radiology and Imaging
- Artificial Intelligence in Medicine
- 3D Imaging Analysis
Background:
- Cone-beam computed tomography (CBCT) is crucial for craniofacial diagnostics.
- Accurate landmark identification is essential for precise 3D measurements in CBCT.
- Manual landmarking is time-consuming and subject to inter-observer variability.
Purpose of the Study:
- To evaluate the clinical validity and accuracy of a deep learning-based automatic landmarking algorithm for CBCT.
- To compare 3D CBCT head measurements derived from manual versus automatic landmark identification.
Main Methods:
- Eighty CBCT scans were analyzed, categorized into non-surgical, surgical without hardware, and surgical with hardware groups.
- Sixty-five landmarks were identified manually and via a 3D automatic landmark detection method.
- Fifty-three measurements (lengths, angles, ratios) were calculated from identified landmarks.
Main Results:
- Deep learning-based automatic landmarking demonstrated accuracy comparable to manual methods.
- Six specific measurements showed statistically significant differences between manual and AI landmarking (P<0.05).
- Automatic landmarking reduced identification time from 40-60 minutes to approximately 10.9 seconds per CBCT volume.
Conclusions:
- The deep learning algorithm for CBCT automatic landmarking is clinically valid and accurate.
- AI-driven landmarking significantly enhances diagnostic and treatment planning efficiency by reducing measurement time.
