MR Datasets for Compressed Sensing
Compressed sensing techniques (sampling below the Nyquist rate) are a popular research focus in magnetic resonance imaging today, because scanner acquisition times speed up dramatically. Parallel imaging, the use of multiple receive surface coils in parallel, complements compressed sensing by providing wide fields of view and/or increased signal-to-noise ratios. The tradeoff in using these techniques is that the image reconstruction step becomes computationally intensive, with a wide variety of noise-like artifcacts arising when pushing the limits of undersamping acceleration.
New reconstruction algorithms for multi-CPU and/or GPU-based servers will speed up image reconstruction and enable compressed sensing to be brought into wide clinical application. This web site provides open datasets to researchers who desire to contribute to a community of reproducible research, where they can test and validate their algorithms against known undersampled acquisitions. These datasets were acquired through a collaboration between Prof. Michael Lustig at UC Berkeley and Dr. Shreyas Vasanawala at Stanford's Lucille Packard Children's Hospital. The undersampled datasets are of two varieties: variable-density undersampling and uniform-density undersampling. At present, all of the datasets are of knee images. In addition to undersampled datasets, we also provide separate cases of fully sampled knees, for researchers who wish to experiment with their own undersampling patterns.