Augmentation based unsupervised domain adaptation

TL;DR

This paper introduces an augmentation-based unsupervised domain adaptation method that enhances the generalization of medical image segmentation models across different data sources, addressing data scarcity and variability issues.

Contribution

It proposes a novel combination of adversarial domain adaptation and consistency training to improve segmentation model robustness in medical imaging.

Findings

Improved model generalization across different datasets.

Enhanced segmentation accuracy in challenging cases.

Effective adaptation with limited annotated data.

Abstract

The insertion of deep learning in medical image analysis had lead to the development of state-of-the art strategies in several applications such a disease classification, as well as abnormality detection and segmentation. However, even the most advanced methods require a huge and diverse amount of data to generalize. Because in realistic clinical scenarios, data acquisition and annotation is expensive, deep learning models trained on small and unrepresentative data tend to outperform when deployed in data that differs from the one used for training (e.g data from different scanners). In this work, we proposed a domain adaptation methodology to alleviate this problem in segmentation models. Our approach takes advantage of the properties of adversarial domain adaptation and consistency training to achieve more robust adaptation. Using two datasets with white matter hyperintensities (WMH) annotations, we demonstrated that the proposed method improves model generalization even in corner cases where individual strategies tend to fail.