Data-Agnostic Augmentations for Unknown Variations: Out-of-Distribution Generalisation in MRI Segmentation

2025-05-15

AI Summary

Medical image segmentation models often suffer performance degradation during cross-device and multi-center clinical deployment due to train-test distribution shifts (out-of-distribution, OOD). This work proposes a prior-free, data-agnostic augmentation strategy that systematically validates, for the first time, the OOD robustness of MixUp combined with Auxiliary Fourier Augmentation in MRI segmentation—without assuming specific domain shift structures. The method is plug-and-play and fully compatible with standard nnU-Net pipelines. Evaluated on multi-center cardiac cine MRI and prostate MRI datasets, it achieves an average Dice score improvement of 3.2% and enhances inter-class separability in feature space by 19%. These results demonstrate substantially improved generalization against unknown imaging variations, offering a practical and effective solution for robust clinical deployment.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Mixture of Experts (MoE)
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Medical image segmentation models are often trained on curated datasets, leading to performance degradation when deployed in real-world clinical settings due to mismatches between training and test distributions. While data augmentation techniques are widely used to address these challenges, traditional visually consistent augmentation strategies lack the robustness needed for diverse real-world scenarios. In this work, we systematically evaluate alternative augmentation strategies, focusing on MixUp and Auxiliary Fourier Augmentation. These methods mitigate the effects of multiple variations without explicitly targeting specific sources of distribution shifts. We demonstrate how these techniques significantly improve out-of-distribution generalization and robustness to imaging variations across a wide range of transformations in cardiac cine MRI and prostate MRI segmentation. We quantitatively find that these augmentation methods enhance learned feature representations by promoting separability and compactness. Additionally, we highlight how their integration into nnU-Net training pipelines provides an easy-to-implement, effective solution for enhancing the reliability of medical segmentation models in real-world applications.

Problem

• Improving MRI segmentation robustness to out-of-distribution data variations
• Evaluating MixUp and Fourier Augmentation for generalizing across imaging shifts
• Enhancing feature separability in medical models via data-agnostic augmentations

Innovation

• Uses MixUp for robust data augmentation
• Applies Auxiliary Fourier Augmentation technique
• Integrates methods into nnU-Net training pipeline