Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cephalometric Landmark Detection across Ages with Prototypical Network

Han Wu, Chong Wang, Lanzhuju Mei, Tong Yang, Min Zhu, Dingggang Shen, Zhiming Cui

TL;DR

CeLDA addresses the need for unified cephalometric landmark detection across adolescents and adults by modeling landmarks with holistic prototypes learned through a prototypical network. It introduces cross-age robustness via exponential-moving prototype estimation, cross-image prototype alignment, and a masked prototype relation mining module that leverages anatomical landmark relations. The method achieves state-of-the-art performance on a newly released CephAdoAdu benchmark, with strong gains for adolescent cases and across-age scenarios, and is supported by comprehensive ablations validating each component. The work also provides the first age-inclusive dataset for cephalometric landmark detection, enabling broader evaluation and future research in clinically realistic settings.

Abstract

Automated cephalometric landmark detection is crucial in real-world orthodontic diagnosis. Current studies mainly focus on only adult subjects, neglecting the clinically crucial scenario presented by adolescents whose landmarks often exhibit significantly different appearances compared to adults. Hence, an open question arises about how to develop a unified and effective detection algorithm across various age groups, including adolescents and adults. In this paper, we propose CeLDA, the first work for Cephalometric Landmark Detection across Ages. Our method leverages a prototypical network for landmark detection by comparing image features with landmark prototypes. To tackle the appearance discrepancy of landmarks between age groups, we design new strategies for CeLDA to improve prototype alignment and obtain a holistic estimation of landmark prototypes from a large set of training images. Moreover, a novel prototype relation mining paradigm is introduced to exploit the anatomical relations between the landmark prototypes. Extensive experiments validate the superiority of CeLDA in detecting cephalometric landmarks on both adult and adolescent subjects. To our knowledge, this is the first effort toward developing a unified solution and dataset for cephalometric landmark detection across age groups. Our code and dataset will be made public on https://github.com/ShanghaiTech-IMPACT/Cephalometric-Landmark-Detection-across-Ages-with-Prototypical-Network

Cephalometric Landmark Detection across Ages with Prototypical Network

TL;DR

CeLDA addresses the need for unified cephalometric landmark detection across adolescents and adults by modeling landmarks with holistic prototypes learned through a prototypical network. It introduces cross-age robustness via exponential-moving prototype estimation, cross-image prototype alignment, and a masked prototype relation mining module that leverages anatomical landmark relations. The method achieves state-of-the-art performance on a newly released CephAdoAdu benchmark, with strong gains for adolescent cases and across-age scenarios, and is supported by comprehensive ablations validating each component. The work also provides the first age-inclusive dataset for cephalometric landmark detection, enabling broader evaluation and future research in clinically realistic settings.

Abstract

Automated cephalometric landmark detection is crucial in real-world orthodontic diagnosis. Current studies mainly focus on only adult subjects, neglecting the clinically crucial scenario presented by adolescents whose landmarks often exhibit significantly different appearances compared to adults. Hence, an open question arises about how to develop a unified and effective detection algorithm across various age groups, including adolescents and adults. In this paper, we propose CeLDA, the first work for Cephalometric Landmark Detection across Ages. Our method leverages a prototypical network for landmark detection by comparing image features with landmark prototypes. To tackle the appearance discrepancy of landmarks between age groups, we design new strategies for CeLDA to improve prototype alignment and obtain a holistic estimation of landmark prototypes from a large set of training images. Moreover, a novel prototype relation mining paradigm is introduced to exploit the anatomical relations between the landmark prototypes. Extensive experiments validate the superiority of CeLDA in detecting cephalometric landmarks on both adult and adolescent subjects. To our knowledge, this is the first effort toward developing a unified solution and dataset for cephalometric landmark detection across age groups. Our code and dataset will be made public on https://github.com/ShanghaiTech-IMPACT/Cephalometric-Landmark-Detection-across-Ages-with-Prototypical-Network
Paper Structure (15 sections, 8 equations, 3 figures, 2 tables)

This paper contains 15 sections, 8 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Overview
  4. Holistic Estimation of Landmark Prototypes
  5. Cross-image Prototype Alignment
  6. Masked Prototype Relation Mining
  7. Overall Training Objective
  8. Experiments
  9. Dataset and Evaluation Metric
  10. Dataset:
  11. Evaluation Metric:
  12. Implementation Details
  13. Comparison with SOTA Approaches
  14. Analytical Ablation Studies
  15. Conclusion

Figures (3)

  • Figure 1: (a) An adult case, with regular anatomical structures and permanent teeth (orange arrow); (b,c,d) adolescent cases, with complicated anatomical changes due to unerupted teeth (blue arrow) and baby teeth (green arrow). These changes on adolescent cases cause significant landmark shifts. Here we show only two landmarks (red points) out of ten for better visualization.
  • Figure 2: An overview of the proposed CeLDA method for cephalometric landmark detection, based on a set of holistic landmark prototypes.
  • Figure 3: (a) Visual comparison of different landmark detection methods, where the ground truth and predicted landmarks are indicated in green and red dots respectively, and each pair of them is linked with a yellow line. (b) Effect of the mask ratio $R$ for mining landmark prototype relations.