Abstract:Accurate and robust medical image segmentation is fundamental and crucial for enhancing the autonomy of computer-aided diagnosis and intervention systems. Medical data collection normally involves different scanners, protocols, and populations, making domain adaptation (DA) a highly demanding research field to alleviate model degradation in the deployment site. To preserve the model performance across multiple testing domains, this work proposes the Curriculum-based Augmented Fourier Domain Adaptation (Curri-AFDA) for robust medical image segmentation. In particular, our curriculum learning strategy is based on the causal relationship of a model under different levels of data shift in the deployment phase, where the higher the shift is, the harder to recognize the variance. Considering this, we progressively introduce more amplitude information from the target domain to the source domain in the frequency space during the curriculum-style training to smoothly schedule the semantic knowledge transfer in an easier-to-harder manner. Besides, we incorporate the training-time chained augmentation mixing to help expand the data distributions while preserving the domain-invariant semantics, which is beneficial for the acquired model to be more robust and generalize better to unseen domains. Extensive experiments on two segmentation tasks of Retina and Nuclei collected from multiple sites and scanners suggest that our proposed method yields superior adaptation and generalization performance. Meanwhile, our approach proves to be more robust under various corruption types and increasing severity levels. In addition, we show our method is also beneficial in the domain-adaptive classification task with skin lesion datasets. The code is available at https://github.com/lofrienger/Curri-AFDA. Note to Practitioners—Medical image segmentation is key to improving computer-assisted diagnosis and intervention autonomy. However, due to domain gaps between different medical sites, deep learning-based segmentation models frequently encounter performance degradation when deployed in a novel domain. Moreover, model robustness is also highly expected to mitigate the effects of data corruption. Considering all these demanding yet practical needs to automate medical applications and benefit healthcare, we propose the Curriculum-based Fourier Domain Adaptation (Curri-AFDA) for medical image segmentation. Extensive experiments on two segmentation tasks with cross-domain datasets show the consistent superiority of our method regarding adaptation and generalization on multiple testing domains and robustness against synthetic corrupted data. Besides, our approach is independent of image modalities because its efficacy does not rely on modality-specific characteristics. In addition, we demonstrate the benefit of our method for image classification besides segmentation in the ablation study. Therefore, our method can potentially be applied in many medical applications and yield improved performance. Future works may be extended by exploring the integration of curriculum learning regime with Fourier domain amplitude fusion in the testing time rather than in the training time like this work and most other existing domain adaptation works.

Domain Adaptation for Medical Image Classification Without Source Data

A Survey on Domain Generalization for Medical Image Analysis

Domain Adaptation Meets Zero-Shot Learning: an Annotation-Efficient Approach to Multi-Modality Medical Image Segmentation

A source free domain adaptation model based on adversarial learning for image classification

Embracing the disharmony in medical imaging: A Simple and effective framework for domain adaptation

Domain Adaptation for the Segmentation of Confidential Medical Images

Collaborative Unsupervised Domain Adaptation for Medical Image Diagnosis

Data privacy protection domain adaptation by roughing and finishing stage

Cross-domain additive learning of new knowledge rather than replacement

Unsupervised Cross-Modality Domain Adaptation of ConvNets for Biomedical Image Segmentations with Adversarial Loss

Confident Learning-Based Domain Adaptation for Hyperspectral Image Classification

Data Augmentation-Based Unsupervised Domain Adaptation In Medical Imaging

Causality-Inspired Single-Source Domain Generalization for Medical Image Segmentation

Curriculum-Based Augmented Fourier Domain Adaptation for Robust Medical Image Segmentation

Unsupervised Domain Adaptation for Mammogram Image Classification: A Promising Tool for Model Generalization

Grade-Skewed Domain Adaptation via Asymmetric Bi-Classifier Discrepancy Minimization for Diabetic Retinopathy Grading

Superpixel-guided class-level denoising for unsupervised domain adaptive fundus image segmentation without source data

Towards Robust Cross-domain Image Understanding with Unsupervised Noise Removal

Unsupervised Domain Adaptation Network with Category-Centric Prototype Aligner for Biomedical Image Segmentation

Towards collaborative appearance and semantic adaptation for medical image segmentation

Domain-aware Dual Attention for Generalized Medical Image Segmentation on Unseen Domains