Towards Ultrafast MRI via Extreme k-Space Undersampling and Superresolution
This work addresses the need for faster MRI imaging for clinical applications like surgery planning, though it is incremental as it builds on existing fastMRI benchmarks and deep learning methods.
The paper tackled the problem of accelerating MRI scans via extreme k-space undersampling and deep learning-based superresolution, achieving an MSE of 0.00114, PSNR of 29.6 dB, and SSIM of 0.956 at x16 acceleration, with diagnostic value preserved in expert assessments.
We went below the MRI acceleration factors (a.k.a., k-space undersampling) reported by all published papers that reference the original fastMRI challenge, and then considered powerful deep learning based image enhancement methods to compensate for the underresolved images. We thoroughly study the influence of the sampling patterns, the undersampling and the downscaling factors, as well as the recovery models on the final image quality for both the brain and the knee fastMRI benchmarks. The quality of the reconstructed images surpasses that of the other methods, yielding an MSE of 0.00114, a PSNR of 29.6 dB, and an SSIM of 0.956 at x16 acceleration factor. More extreme undersampling factors of x32 and x64 are also investigated, holding promise for certain clinical applications such as computer-assisted surgery or radiation planning. We survey 5 expert radiologists to assess 100 pairs of images and show that the recovered undersampled images statistically preserve their diagnostic value.