AADG: Automatic Augmentation for Domain Generalization on Retinal Image Segmentation

TL;DR

This paper introduces AADG, an automated data augmentation method that enhances domain generalization in retinal image segmentation by diversifying training data through optimized policies, leading to state-of-the-art results across multiple datasets.

Contribution

The paper presents a novel automated augmentation framework using adversarial training and reinforcement learning to improve domain generalization in medical image segmentation.

Findings

Achieves state-of-the-art performance on multiple retinal segmentation datasets.

Demonstrates effective cross-modality generalization on OCTA datasets.

Proves the learned policies are model-agnostic and transferable.

Abstract

Convolutional neural networks have been widely applied to medical image segmentation and have achieved considerable performance. However, the performance may be significantly affected by the domain gap between training data (source domain) and testing data (target domain). To address this issue, we propose a data manipulation based domain generalization method, called Automated Augmentation for Domain Generalization (AADG). Our AADG framework can effectively sample data augmentation policies that generate novel domains and diversify the training set from an appropriate search space. Specifically, we introduce a novel proxy task maximizing the diversity among multiple augmented novel domains as measured by the Sinkhorn distance in a unit sphere space, making automated augmentation tractable. Adversarial training and deep reinforcement learning are employed to efficiently search the objectives. Quantitative and qualitative experiments on 11 publicly-accessible fundus image datasets (four for retinal vessel segmentation, four for optic disc and cup (OD/OC) segmentation and three for retinal lesion segmentation) are comprehensively performed. Two OCTA datasets for retinal vasculature segmentation are further involved to validate cross-modality generalization. Our proposed AADG exhibits state-of-the-art generalization performance and outperforms existing approaches by considerable margins on retinal vessel, OD/OC and lesion segmentation tasks. The learned policies are empirically validated to be model-agnostic and can transfer well to other models. The source code is available at https://github.com/CRazorback/AADG.