AADG: Automatic Augmentation for Domain Generalization on Retinal Image Segmentation

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.

arXiv commentAccepted by IEEE Transactions on Medical Imaging (TMI)

References in corpus (17)

• mixup: Beyond Empirical Risk Minimization
• CE-Net: Context Encoder Network for 2D Medical Image Segmentation
• EfficientNetV2: Smaller Models and Faster Training
• REFUGE Challenge: A Unified Framework for Evaluating Automated Methods for Glaucoma Assessment from Fundus Photographs
• Joint Optic Disc and Cup Segmentation Based on Multi-label Deep Network and Polar Transformation
• Domain Generalization via Model-Agnostic Learning of Semantic Features
• Generalizing to Unseen Domains via Adversarial Data Augmentation
• Maximum Density Divergence for Domain Adaptation
• DoFE: Domain-oriented Feature Embedding for Generalizable Fundus Image Segmentation on Unseen Datasets
• Learning to Compose Domain-Specific Transformations for Data Augmentation
• Domain Agnostic Learning with Disentangled Representations
• Improving the generalizability of convolutional neural network-based segmentation on CMR images
• Generalizing to Unseen Domains: A Survey on Domain Generalization
• Adversarial AutoAugment
• DomainMix: Learning Generalizable Person Re-Identification Without Human Annotations
• DivAug: Plug-in Automated Data Augmentation with Explicit Diversity Maximization
• Student Becomes Decathlon Master in Retinal Vessel Segmentation via Dual-teacher Multi-target Domain Adaptation

Cited by in corpus (3)

• Domain Generalization for Medical Image Analysis: A Review
• Segmenting Medical Images with Limited Data
• MedAugment: Universal Automatic Data Augmentation Plug-in for Medical Image Analysis