DRL-STNet: Unsupervised Domain Adaptation for Cross-modality Medical Image Segmentation via Disentangled Representation Learning

TL;DR

DRL-STNet introduces an unsupervised domain adaptation framework for cross-modality medical image segmentation, combining GANs, disentangled representation learning, and self-training to improve accuracy and reduce manual annotation dependency.

Contribution

The paper proposes a novel DRL-STNet framework that effectively translates images between modalities and iteratively refines segmentation models using pseudo-labels, outperforming existing methods.

Findings

Achieved 11.4% higher Dice score than state-of-the-art.

Surpassed baseline in Normalized Surface Dice by 13.1%.

Demonstrated efficient training with 41 seconds runtime.

Abstract

Unsupervised domain adaptation (UDA) is essential for medical image segmentation, especially in cross-modality data scenarios. UDA aims to transfer knowledge from a labeled source domain to an unlabeled target domain, thereby reducing the dependency on extensive manual annotations. This paper presents DRL-STNet, a novel framework for cross-modality medical image segmentation that leverages generative adversarial networks (GANs), disentangled representation learning (DRL), and self-training (ST). Our method leverages DRL within a GAN to translate images from the source to the target modality. Then, the segmentation model is initially trained with these translated images and corresponding source labels and then fine-tuned iteratively using a combination of synthetic and real images with pseudo-labels and real labels. The proposed framework exhibits superior performance in abdominal organ segmentation on the FLARE challenge dataset, surpassing state-of-the-art methods by 11.4% in the Dice similarity coefficient and by 13.1% in the Normalized Surface Dice metric, achieving scores of 74.21% and 80.69%, respectively. The average running time is 41 seconds, and the area under the GPU memory-time curve is 11,292 MB. These results indicate the potential of DRL-STNet for enhancing cross-modality medical image segmentation tasks.