Cooperative Training and Latent Space Data Augmentation for Robust Medical Image Segmentation
This work addresses robustness issues in medical image segmentation for deployment in clinical settings, but it is incremental as it builds on existing methods for data augmentation and network cooperation.
The paper tackles the problem of deep learning-based medical image segmentation models being vulnerable to data distribution shifts, such as different scanners or imaging artifacts, by proposing a cooperative training framework and latent space data augmentation method, resulting in a 15% improvement in average Dice score compared to standard training with only 10 subjects from a single site.
Deep learning-based segmentation methods are vulnerable to unforeseen data distribution shifts during deployment, e.g. change of image appearances or contrasts caused by different scanners, unexpected imaging artifacts etc. In this paper, we present a cooperative framework for training image segmentation models and a latent space augmentation method for generating hard examples. Both contributions improve model generalization and robustness with limited data. The cooperative training framework consists of a fast-thinking network (FTN) and a slow-thinking network (STN). The FTN learns decoupled image features and shape features for image reconstruction and segmentation tasks. The STN learns shape priors for segmentation correction and refinement. The two networks are trained in a cooperative manner. The latent space augmentation generates challenging examples for training by masking the decoupled latent space in both channel-wise and spatial-wise manners. We performed extensive experiments on public cardiac imaging datasets. Using only 10 subjects from a single site for training, we demonstrated improved cross-site segmentation performance and increased robustness against various unforeseen imaging artifacts compared to strong baseline methods. Particularly, cooperative training with latent space data augmentation yields 15% improvement in terms of average Dice score when compared to a standard training method.