Edge-preserving Domain Adaptation for semantic segmentation of Medical Images

TL;DR

This paper introduces an edge-preserving unsupervised domain adaptation method for medical image segmentation, improving the retention of fine details like blood vessels during domain shifts.

Contribution

It proposes a novel cycle-consistent loss combined with an edge-based loss to maintain high-level semantic details during domain adaptation.

Findings

Achieves 1.1 to 9.2 higher DICE scores than state-of-the-art methods.

Outperforms vanilla CycleGAN by approximately 5.2 DICE points.

Effective in preserving fine structures like blood vessels in medical images.

Abstract

Domain Adaptation is a technique to address the lack of massive amounts of labeled data in unseen environments. Unsupervised domain adaptation is proposed to adapt a model to new modalities using solely labeled source data and unlabeled target domain data. Though many image-spaces domain adaptation methods have been proposed to capture pixel-level domain-shift, such techniques may fail to maintain high-level semantic information for the segmentation task. For the case of biomedical images, fine details such as blood vessels can be lost during the image transformation operations between domains. In this work, we propose a model that adapts between domains using cycle-consistent loss while maintaining edge details of the original images by enforcing an edge-based loss during the adaptation process. We demonstrate the effectiveness of our algorithm by comparing it to other approaches on two eye fundus vessels segmentation datasets. We achieve 1.1 to 9.2 increment in DICE score compared to the SOTA and ~5.2 increments compared to a vanilla CycleGAN implementation.