Detecting Dental Landmarks from Intraoral 3D Scans: the 3DTeethLand challenge

2025-12-09

AI Summary

This study addresses the challenge of automatic 3D intraoral tooth landmark detection. To overcome complexities arising from anatomical variability and morphological diversity, we introduce a deep learning–based framework and publicly release the first large-scale, expert-annotated 3D dental landmark dataset—filling a critical gap in benchmark resources. Leveraging the MICCAI International Challenge, we establish a standardized evaluation platform to foster algorithmic innovation. Our method unifies processing of both point cloud and mesh inputs, enabling sub-millimeter localization of key anatomical landmarks. Extensive experiments demonstrate significant improvements in detection accuracy and robustness across diverse clinical scenarios; multiple participating algorithms achieve clinically viable performance. The proposed framework provides a reproducible, rigorously validated foundation for intelligent orthodontic diagnosis, personalized treatment planning, and dynamic therapeutic assessment.

Technology Category

• Computer Vision: 3D Computer Vision
• Intelligent Robots: Localization, Mapping, and Navigation
• Machine Learning: Deep Neural Architectures and Foundation Models

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Teeth landmark detection is a critical task in modern clinical orthodontics. Their precise identification enables advanced diagnostics, facilitates personalized treatment strategies, and supports more effective monitoring of treatment progress in clinical dentistry. However, several significant challenges may arise due to the intricate geometry of individual teeth and the substantial variations observed across different individuals. To address these complexities, the development of advanced techniques, especially through the application of deep learning, is essential for the precise and reliable detection of 3D tooth landmarks. In this context, the 3DTeethLand challenge was held in collaboration with the International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) in 2024, calling for algorithms focused on teeth landmark detection from intraoral 3D scans. This challenge introduced the first publicly available dataset for 3D teeth landmark detection, offering a valuable resource to assess the state-of-the-art methods in this task and encourage the community to provide methodological contributions towards the resolution of their problem with significant clinical implications.

Problem

• Detecting dental landmarks from intraoral 3D scans for orthodontics.
• Addressing challenges from intricate tooth geometry and individual variations.
• Developing deep learning techniques for precise 3D teeth landmark detection.

Innovation

• Deep learning for 3D tooth landmark detection
• First public dataset for 3D teeth landmark detection
• Challenge to develop algorithms for intraoral 3D scans