Eigen-reconstructions: a closer look into the System Matrix

Abstract

Magnetic particle imaging (MPI) offers an exceptional set of advantages including high sampling efficiency and sub-millimeter spatial resolution. The former is maximized using ad-hoc sampling trajectories; the latter relies on the compensation of the point spread function of the superparamagnetic iron oxide particles (SPIOs) necessary for the MPI signal. The System Matrix (SM) approach for reconstruction uses a pre-calibration measurement and achieves both purposes simultaneously, relating the concentration of SPIOs and the particle response to the signal. Consequently, the SM’s quality will largely influence the reconstruction. Considering the multitude of factors involved in the reconstruction, it is difficult to identify sources of image artifacts.  In this work, we demonstrate the potential to use reconstructions of individual measurements within the SM (eigen-reconstruction) as a test for both SM and reconstruction quality. We also present an algorithm to enhance image quality using eigen-reconstructions.

Int. J. Mag. Part. Imag. 6(2), Suppl. 1, 2020, Article ID: 2009072, DOI: 10.18416/IJMPI.2020.2009072