Improving Artifact Robustness for CT Deep Learning Models Without Labeled Artifact Images via Domain Adaptation
This addresses robustness issues in medical imaging for clinical deployment where novel artifacts can emerge, offering a cost-effective alternative to labeling, though it is incremental as it applies existing domain adaptation methods to a specific domain.
The study tackled the problem of CT deep learning models degrading when encountering new scanner artifacts by using domain adaptation to maintain classification performance without labeled artifact images, achieving comparable accuracy to models trained with labeled artifacts and showing unexpected generalization to uniform noise.
Deep learning models which perform well on images from their training distribution can degrade substantially when applied to new distributions. If a CT scanner introduces a new artifact not present in the training labels, the model may misclassify the images. Although modern CT scanners include design features which mitigate these artifacts, unanticipated or difficult-to-mitigate artifacts can still appear in practice. The direct solution of labeling images from this new distribution can be costly. As a more accessible alternative, this study evaluates domain adaptation as an approach for training models that maintain classification performance despite new artifacts, even without corresponding labels. We simulate ring artifacts from detector gain error in sinogram space and evaluate domain adversarial neural networks (DANN) against baseline and augmentation-based approaches on the OrganAMNIST abdominal CT dataset. Our results demonstrate that baseline models trained only on clean images fail to generalize to images with ring artifacts, and traditional augmentation with other distortion types provides no improvement on unseen artifact domains. In contrast, the DANN approach successfully maintains high classification accuracy on ring artifact images using only unlabeled artifact data during training, demonstrating the viability of domain adaptation for artifact robustness. The domain-adapted model achieved classification performance on ring artifact test data comparable to models explicitly trained with labeled artifact images, while also showing unexpected generalization to uniform noise. These findings provide empirical evidence that domain adaptation can effectively address distribution shift in medical imaging without requiring expensive expert labeling of new artifact distributions, suggesting promise for deployment in clinical settings where novel artifacts may emerge.