A main challenge in Magnetic Resonance Imaging (MRI) for clinical applications is speeding up scan time. Beyond the improvement of patient experience and the reduction of operational costs, faster scans are essential for time-sensitive imaging, where target movement is unavoidable, yet must be significantly lessened, e.g., fetal MRI, cardiac cine, and lungs imaging. Moreover, short scan time can enhance temporal resolution in dynamic scans, such as functional MRI or dynamic contrast enhanced MRI. Current imaging methods facilitate MRI acquisition at the price of lower spatial resolution and costly hardware solutions.
We introduce a practical, software-only framework, based on deep learning, for accelerating MRI acquisition, while maintaining anatomically meaningful imaging. This is accomplished by partial MRI sampling, while using an adversarial neural network to directly estimate the missing k-space samples. The inter-play between the generator and the discriminator networks enables the introduction of an adversarial cost in addition to a fidelity loss used for optimizing the peak signal-to-noise ratio (PSNR). Promising image reconstruction results are obtained for 3T and 1.5T brain MRI, from large publicly available dataset, where only 40%, 25% and 16.6% of the raw samples of each scan are used. To assess the clinical usability of the reconstructed images we also performed tissue segmentation and compared the results to those obtained by using the original fully-sampled images.
Segmentation compatibility, measured in terms of Dice scores and Hausdorff distances, demonstrate the quality of the proposed MRI reconstruction with respect to other methods, including the widely-used Compressed Sensing.