Motion Artifact Reduction in Quantitative Susceptibility Mapping using Deep Neural Network
This addresses motion artifact reduction in medical imaging for applications like Parkinson's disease diagnosis, but appears incremental as it applies existing deep learning methods to a specific domain problem.
The paper tackled motion artifacts in Quantitative Susceptibility Mapping by proposing a deep learning approach that uses simulated motion-corrupted images for training, and demonstrated successful suppression of artifacts like ringing and ghosting in tests on simulated data, healthy volunteers, and Parkinson's disease patients.
An approach to reduce motion artifacts in Quantitative Susceptibility Mapping using deep learning is proposed. We use an affine motion model with randomly created motion profiles to simulate motion-corrupted QSM images. The simulated QSM image is paired with its motion-free reference to train a neural network using supervised learning. The trained network is tested on unseen simulated motion-corrupted QSM images, in healthy volunteers and in Parkinson's disease patients. The results show that motion artifacts, such as ringing and ghosting, were successfully suppressed.