CDPM-Align: Multi-Scale Guidance-Aligned Diffusion Pretraining for Robust Few-Shot Anatomical Landmark Detection

2026-06-03

AI Summary

This work addresses the limited robustness and reliability of anatomical landmark detection in medical images under few-shot and low-annotation settings. To tackle this challenge, the authors propose a conditional diffusion-based pretraining approach with multi-scale guided alignment. This method introduces, for the first time, a multi-scale feature alignment mechanism into diffusion models to enhance representation learning for anatomical landmarks. By performing generative pretraining on a small set of heterogeneous medical images, the model substantially improves performance on downstream few-shot detection tasks. Experimental results demonstrate that, with only 10–25 annotated images, the proposed approach consistently achieves higher localization accuracy and prediction reliability across multiple benchmark datasets, offering strong support for safe clinical deployment.

Abstract

Anatomical landmark detection is a fundamental task in medical image analysis supporting a wide range of diagnostic and interventional workflows. Although recent methods have achieved sub-millimetric localisation, accuracy alone is not sufficient for clinical deployment, requiring reliability and robustness in prediction. Despite its clinical relevance, the impact of representation learning in this context is still underexplored. In this work, we introduce CDPM-align, a multi-scale guidance-aligned conditional diffusion pre-training for anatomical landmark detection. Our experimental setup focuses on a few images and a few annotation regimes. Specifically, we employ three popular heterogeneous small-scale benchmark datasets for representation learning via conditional generative pre-training. Furthermore, we consider low-annotation scenarios for the downstream task of landmark detection, with 10 and 25 annotated images, reflecting realistic trade-offs between clinical effort and resource constraints for annotations. Our results confirm that generative pre-training enables the model to learn a robust representation. This improves both accuracy and uncertainty on the downstream tasks, advancing towards safe and efficient clinical deployment.

Problem

• anatomical landmark detection
• few-shot learning
• representation learning
• robustness
• medical image analysis

Innovation

• diffusion pretraining
• few-shot landmark detection
• multi-scale guidance
• medical image analysis
• robust representation learning