Reducing Domain Gap in Frequency and Spatial domain for Cross-modality Domain Adaptation on Medical Image Segmentation

TL;DR

This paper introduces a simple frequency and spatial domain transfer method within a multi-teacher distillation framework to improve unsupervised domain adaptation for medical image segmentation, avoiding complex adversarial training.

Contribution

It proposes a novel frequency and spatial domain transfer approach combined with multi-teacher distillation for effective UDA in medical imaging, bypassing adversarial learning.

Findings

Achieves superior performance on cross-modality medical segmentation datasets.

Effectively narrows domain gap using frequency domain component replacement.

Reduces style bias with a batch momentum histogram matching strategy.

Abstract

Unsupervised domain adaptation (UDA) aims to learn a model trained on source domain and performs well on unlabeled target domain. In medical image segmentation field, most existing UDA methods depend on adversarial learning to address the domain gap between different image modalities, which is ineffective due to its complicated training process. In this paper, we propose a simple yet effective UDA method based on frequency and spatial domain transfer uner multi-teacher distillation framework. In the frequency domain, we first introduce non-subsampled contourlet transform for identifying domain-invariant and domain-variant frequency components (DIFs and DVFs), and then keep the DIFs unchanged while replacing the DVFs of the source domain images with that of the target domain images to narrow the domain gap. In the spatial domain, we propose a batch momentum update-based histogram matching strategy to reduce the domain-variant image style bias. Experiments on two cross-modality medical image segmentation datasets (cardiac, abdominal) show that our proposed method achieves superior performance compared to state-of-the-art methods.