Point-of-care magnetic resonance imaging (MRI) requires clear images within a short scanning time, a small footprint of the scanner, and relatively low memory required for image reconstruction. A permanent magnet array (PMA)-based MRI system is a good candidate to supply a magnetic field due to its compactness and low power consumption. However, it has relatively inhomogeneous magnetic field and thus non-linear gradients, which results in location-dependent k-spaces (so called local k-spaces) and uneven signal point populations in the local k-spaces, compromising the image quality. Moreover, owing to the non-linearity, imaging reconstruction using Fourier transform does not work, which leads to an increase in the required computation memory. In this study, in order to improve the image quality, the approaches of compensating the uneven signal point population by increasing the numbers of sampling points or rotation angles are investigated in terms of their impacts on image quality improvement, acquisition time, image reconstruction time, and memory consumption. Both methods give a significant improvement on image image quality although they result in a large and dense encoding matrix and thus a large memory consumption. To lower the memory consumption, it is further proposed to transform such a matrix to frequency domain where the matrix could be sparse. Moreover, a row-wise truncation to the transformed encoding matrix is applied to further reduce the memory consumption. Through the results of numerical experiments, it is shown that the required memory for calculation can effectively be reduced by 71.6% while the image becomes clearer by increasing the number of sampling point and/or the number of rotation angles. With the successful demonstration where improved image quality and a lowered memory required can be obtained simultaneously, the proposed study is one step forward for a PMA-based MRI system towards its targeted point-of-care application scenario.
This is an open access article distributed under the Creative Commons Attribution License
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited