DDFP: Data-dependent Frequency Prompt for Source Free Domain Adaptation of Medical Image Segmentation

TL;DR

This paper introduces a novel source-free domain adaptation framework for medical image segmentation that uses data-dependent frequency prompts and style-related layer fine-tuning to improve pseudo-label quality and adaptation efficiency.

Contribution

It proposes a preadaptation step for better initialization, a data-dependent frequency prompt for style translation, and a style-related layer fine-tuning strategy, advancing source-free domain adaptation methods.

Findings

Outperforms existing state-of-the-art methods on abdominal and cardiac segmentation tasks.

Effectively generates high-quality pseudo-labels without extra parameters.

Enhances adaptation efficiency and segmentation accuracy.

Abstract

Domain adaptation addresses the challenge of model performance degradation caused by domain gaps. In the typical setup for unsupervised domain adaptation, labeled data from a source domain and unlabeled data from a target domain are used to train a target model. However, access to labeled source domain data, particularly in medical datasets, can be restricted due to privacy policies. As a result, research has increasingly shifted to source-free domain adaptation (SFDA), which requires only a pretrained model from the source domain and unlabeled data from the target domain data for adaptation. Existing SFDA methods often rely on domain-specific image style translation and self-supervision techniques to bridge the domain gap and train the target domain model. However, the quality of domain-specific style-translated images and pseudo-labels produced by these methods still leaves room for improvement. Moreover, training the entire model during adaptation can be inefficient under limited supervision. In this paper, we propose a novel SFDA framework to address these challenges. Specifically, to effectively mitigate the impact of domain gap in the initial training phase, we introduce preadaptation to generate a preadapted model, which serves as an initialization of target model and allows for the generation of high-quality enhanced pseudo-labels without introducing extra parameters. Additionally, we propose a data-dependent frequency prompt to more effectively translate target domain images into a source-like style. To further enhance adaptation, we employ a style-related layer fine-tuning strategy, specifically designed for SFDA, to train the target model using the prompted target domain images and pseudo-labels. Extensive experiments on cross-modality abdominal and cardiac SFDA segmentation tasks demonstrate that our proposed method outperforms existing state-of-the-art methods.