Fast Construction of System Matrix for Narrowband Magnetic Particle Imaging Based on PSF Estimation

Narrowband Magnetic Particle Imaging (MPI) is a promising molecular imaging technique capable of high-sensitivity quantitative in vivo visualization of superparamagnetic iron oxide nanoparticles (SPIONs), offering a reliable pathway for the development of large field-of-view (FOV) MPI scanners for human-scale applications. In narrowband MPI, raw harmonic images are often compromised by negative artifacts, which can be effectively mitigated through system matrix reconstruction to enhance image quality. However, the calibration of the system matrix is an extremely time-consuming process, particularly when the voxel spacing is set to be dense, posing a significant challenge. In this paper, we propose a fast construction of system matrix (FCSM) method based on point spread function (PSF) estimation. The method estimates the distribution of the PSF at various positions within the FOV to simulate the spatial response of the calibration points, and combines the calibration data from all points to build a refined system matrix. We developed a narrowband MPI scanner to validate the feasibility of this approach. Experimental results demonstrate that the method requires only a single PSF acquisition and less than 10 seconds of computation to complete the system matrix construction, achieving higher reconstruction quality.