Day-Night Adaptation: An Innovative Source-free Adaptation Framework for Medical Image Segmentation

TL;DR

This paper introduces DyNA, a novel source-free, test-time adaptation framework for medical image segmentation that leverages day-night cycles and test data reuse to improve model performance without source data access.

Contribution

DyNA is the first to utilize day-night cycles and test data reuse for source-free domain adaptation in medical image segmentation.

Findings

DyNA outperforms existing TTA and SFDA methods on benchmark tasks.

The framework effectively leverages day-night cycles for model adaptation.

Test data reuse enhances model stability and performance.

Abstract

Distribution shifts widely exist in medical images acquired from different medical centres, hindering the deployment of semantic segmentation models trained on one centre (source domain) to another (target domain). While unsupervised domain adaptation has shown significant promise in mitigating these shifts, it poses privacy risks due to sharing data between centres. To facilitate adaptation while preserving data privacy, source-free domain adaptation (SFDA) and test-time adaptation (TTA) have emerged as effective paradigms, relying solely on target domain data. However, SFDA requires a pre-collected target domain dataset before deployment. TTA insufficiently exploit the potential value of test data, as it processes the test data only once. Considering that most medical centres operate during the day and remain inactive at night in clinical practice, we propose a novel adaptation framework called Day-Night Adaptation (DyNA) with above insights, which performs adaptation through day-night cycles without requiring access to source data. During the day, a low-frequency prompt is trained to adapt the frozen model to each test sample. We construct a memory bank for prompt initialization and develop a warm-up mechanism to enhance prompt training. During the night, we reuse test data collected from the day and introduce a global student model to bridge the knowledge between teacher and student models, facilitating model fine-tuning while ensuring training stability. Extensive experiments demonstrate that our DyNA outperforms existing TTA and SFDA methods on two benchmark medical image segmentation tasks. Code will be available after the paper is published.