Automatic Cephalometric Landmark Localization on CBCT-Derived Digitally Reconstructed Radiographs for Skeletal Malocclusion Classification
2026-08-17 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed CephViT, a computer program that automatically finds important points on side-view X-ray images of the head, which usually takes a long time to do by hand. They tested CephViT on standard 2D X-ray images and found it was quite accurate. They also used computer-generated 2D images made from 3D scans and showed that the program could help classify types of jaw alignment problems with similar accuracy to manual methods. This suggests their approach could help analyze 3D scans more easily by using these 2D images. Overall, their work supports automated analysis of head X-rays for dental assessments.
Cephalometric landmark annotationVision TransformerLateral cephalogramMean radial errorSuccessful detection rate3D CBCT scansDigitally reconstructed radiographs (DRRs)Skeletal malocclusion classificationCoordinate normalizationAutomated cephalometric analysis
Authors
Benjamin Hou, Konstantinia Almpani, Janice S. Lee, Zhiyong Lu
Abstract
Manual cephalometric landmark annotation is important for craniofacial assessment but is labor-intensive and difficult to scale. We introduce CephViT, a Vision Transformer-based model for automated 2D lateral cephalometric landmark localization, and evaluate its use in downstream skeletal malocclusion classification. CephViT was trained and benchmarked on a public lateral cephalogram dataset, achieving a mean radial error of 1.28 +/- 1.42 mm and a successful detection rate of 92.0% at 3.0 mm. Because the private evaluation cohort consisted of 3D CBCT scans, lateral cephalogram-like digitally reconstructed radiographs (DRRs) were generated from each volume and used as 2D inputs to the landmark localization model. Landmark coordinates were normalized into a common coordinate frame, and skeletal malocclusion classification was performed using landmarks shared between the reference and DRR-based pipelines. Classification performance using DRR-localized landmarks was comparable to that obtained using manually annotated reference landmarks, with accuracies of 70.0% and 68.3%, respectively. These results support the feasibility of automated cephalometric analysis on CBCT-derived DRRs for skeletal malocclusion assessment.