Counterfactual MRI Data Augmentation using Conditional Denoising Diffusion Generative Models
This addresses domain and covariate shifts in medical imaging, potentially enhancing model performance across diverse imaging conditions, though it appears incremental as it applies a known generative method to a specific bottleneck.
The paper tackles the problem of poor generalizability and robustness in deep learning models for medical imaging due to variations in image acquisition parameters by introducing a method using conditional denoising diffusion generative models to generate counterfactual MRI images for data augmentation, resulting in improved segmentation accuracy in out-of-distribution settings.
Deep learning (DL) models in medical imaging face challenges in generalizability and robustness due to variations in image acquisition parameters (IAP). In this work, we introduce a novel method using conditional denoising diffusion generative models (cDDGMs) to generate counterfactual magnetic resonance (MR) images that simulate different IAP without altering patient anatomy. We demonstrate that using these counterfactual images for data augmentation can improve segmentation accuracy, particularly in out-of-distribution settings, enhancing the overall generalizability and robustness of DL models across diverse imaging conditions. Our approach shows promise in addressing domain and covariate shifts in medical imaging. The code is publicly available at https: //github.com/pedromorao/Counterfactual-MRI-Data-Augmentation