Geometric Deep Learning for Automated Landmarking of Maxillary Arches on 3D Oral Scans from Newborns with Cleft Lip and Palate

2025-01-27

AI Summary

To address the time-consuming and subjective nature of manual landmark annotation on 3D scans of the maxillary dental arch in neonates with cleft lip and palate, this paper proposes the first geometric deep learning framework tailored for small-sample, malformed craniofacial 3D point clouds. Methodologically, it innovatively integrates spherical convolutional neural networks (Spherical CNNs) with multi-scale geometric features to achieve end-to-end automatic landmark localization. Contributions include: (1) the first application of geometric deep learning to neonatal cleft lip and palate 3D landmarking; (2) achieving 94.44% landmark detection accuracy and a mean absolute error of 1.676 ± 0.959 mm using only 100 clinical cases—significantly outperforming existing semi-automatic approaches; and (3) delivering high precision, strong generalizability across anatomical variations, and clinically interpretable predictions, thereby providing a robust technical foundation for personalized treatment planning.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Learning with Manifolds
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Rapid advances in 3D model scanning have enabled the mass digitization of dental clay models. However, most clinicians and researchers continue to use manual morphometric analysis methods on these models such as landmarking. This is a significant step in treatment planning for craniomaxillofacial conditions. We aimed to develop and test a geometric deep learning model that would accurately and reliably label landmarks on a complicated and specialized patient population -- infants, as accurately as a human specialist without a large amount of training data. Our developed pipeline demonstrated an accuracy of 94.44% with an absolute mean error of 1.676 +/- 0.959 mm on a set of 100 models acquired from newborn babies with cleft lip and palate. Our proposed pipeline has the potential to serve as a fast, accurate, and reliable quantifier of maxillary arch morphometric features, as well as an integral step towards a future fully automated dental treatment pipeline.

Problem

• 3D oral imagery
• automated landmark detection
• cleft lip and palate

Innovation

• 3D oral modeling
• automated cleft identification
• digital workflow efficiency