Landmarks Are Alike Yet Distinct: Harnessing Similarity and Individuality for One-Shot Medical Landmark Detection

Xu He, Zhen huang, Qingsong Yao, XiaoqianZhou, S Kevin Zhou

Keywords: Medical landmark detection, One-shot learning

TL;DR: A one-shot landmark detection framework that leverages both anatomical similarity and individuality, validated on dental-related head X-rays.

Abstract: Landmark detection plays a crucial role in medical imaging applications such as disease diagnosis, bone age estimation, and therapy planning. However, training models for detecting multiple landmarks simultaneously often encounters the "seesaw phenomenon", where improvements in detecting certain landmarks lead to declines in detecting others. Yet, training a separate model for each landmark increases memory usage and computational overhead. To address these challenges, we propose a novel approach based on the belief that "landmarks are distinct" by training models with pseudo-labels and template data updated continuously during the training process, where each model is dedicated to detecting a single landmark to achieve high accuracy. Furthermore, grounded on the belief that "landmarks are also alike", we introduce an adapter-based fusion model, combining shared weights with landmark-specific weights, to efficiently share model parameters while allowing flexible adaptation to individual landmarks. This approach not only significantly reduces memory and computational resource requirements but also effectively mitigates the seesaw phenomenon in multi-landmark training. Experimental results on publicly available medical image datasets demonstrate that the single-landmark models significantly outperform traditional multi-point joint training models in detecting individual landmarks. Although our adapter-based fusion model shows slightly lower performance compared to the combined results of all single-landmark models, it still surpasses the current state-of-the-art methods while achieving a notable improvement in resource efficiency.

Changes Summary:

Since the original submission, we have made the following revisions:
1. Updated notation: symbols such as Conv-A_k are now represented using superscripts for clarity.
2. Improved figure presentation: enlarged fonts in Figures 1 and 2, and optimized the layout of Figure 2.
3. Added details on the training procedure of baseline methods.
4. Expanded the discussion on landmarks 10, 15, and 18 to provide clearer explanations.
5. Minor formatting and typesetting improvements for better readability.

Main Tex File: main.tex

Confirm Latex Only: true

Authors Changed: false

Arxiv Update Plans: https://arxiv.org/abs/2503.16058

Submission Number: 9