Adapting to Unseen Vendor Domains for MRI Lesion Segmentation
This addresses the issue of domain shift in medical imaging for researchers and practitioners, but it is incremental as it builds on existing image translation methods.
The paper tackled the problem of poor performance of machine learning models on out-of-domain data in MRI lesion segmentation by using an unsupervised image translation model to create synthetic data, finding that training on synthetic data from labels to images yielded Dice scores of 0.63, 0.64, and 0.58 for different vendors, close to direct training scores of 0.65, 0.72, and 0.61.
One of the key limitations in machine learning models is poor performance on data that is out of the domain of the training distribution. This is especially true for image analysis in magnetic resonance (MR) imaging, as variations in hardware and software create non-standard intensities, contrasts, and noise distributions across scanners. Recently, image translation models have been proposed to augment data across domains to create synthetic data points. In this paper, we investigate the application an unsupervised image translation model to augment MR images from a source dataset to a target dataset. Specifically, we want to evaluate how well these models can create synthetic data points representative of the target dataset through image translation, and to see if a segmentation model trained these synthetic data points would approach the performance of a model trained directly on the target dataset. We consider three configurations of augmentation between datasets consisting of translation between images, between scanner vendors, and from labels to images. It was found that the segmentation models trained on synthetic data from labels to images configuration yielded the closest performance to the segmentation model trained directly on the target dataset. The Dice coeffcient score per each target vendor (GE, Siemens, Philips) for training on synthetic data was 0.63, 0.64, and 0.58, compared to training directly on target dataset was 0.65, 0.72, and 0.61.