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

TL;DR

The paper presents the 3DTeethLand MICCAI 2024 challenge, introducing the first public dataset for 3D teeth landmark detection from intraoral scans and a rigorous evaluation protocol. It surveys prior work, details six top-performing methods (with and without tooth segmentation), and analyzes results using mAP and mAR across four landmark categories under a distance-based hit criterion. A statistically robust ranking framework (bootstrapped Wilcoxon tests) demonstrates that Radboud and ChohoTech achieve the strongest landmark localization performance, while others show varying strengths and weaknesses. The study highlights the practical importance of reliable 3D landmark detection for orthodontic planning and suggests future directions including dataset expansion, more diverse data, and robustness to real-world scanning variations.

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.

Figures (12)

• Figure 1: Overview of tooth structure and anatomical landmarks. (a) Illustration of the human dental arches showing the distribution of incisors, canines, premolars, and molars in the upper and lower jaws. (b) 3D intraoral scans annotated with key dental landmarks, including distal, mesial, facial, cusp, outer, and inner points used for geometric analysis.
• Figure 2: Example of the file structure of landmarks annotation.
• Figure 3: Overview of the proposed method by the Radboud team. The method consists of two stages dedicated to tooth instance segmentation and tooth landmark detection.
• Figure 4: Overview of the proposed method proposed by the YY-LAB team. The presented approach consists of two stages: 3D tooth segmentation and landmark detection.
• Figure 5: Overview of the proposed method proposed by the YN-LAB team: Tooth segmentation and data augmentation are performed as a preliminary step, followed by coarse landmark detection guided by the pseudo ground truth and precise landmark detection.