Conventional parallel imaging methods mostly utilize the spatial encoding by array of receiver coils to unfold the periodic aliasing artifact resulted from uniformly undersampled k-space data. In scenarios such as low- or ultra-low-field MRI where effective receiver arrays do not exist and SNRs are low, these methods are not generally applicable. This study presents a U-Net based deep learning approach to single-channel MRI acceleration that unfolds the aliasing by exploiting its periodicity. The results demonstrate the aliasing unfolding capability of this method for single-channel MRI even at very high acceleration and in presence of pathologies.
This abstract and the presentation materials are available to members only; a login is required.