Projection X-space Image Reconstruction for Open-sided FFL Magnetic Particle Imaging

Abstract

Magnetic particle imaging is a novel tracer-based biomedical imaging modality with high sensitivity and linear quantitative properties. FFL-based magnetic particle imaging improves the sensitivity by an order of magnitude, but the quality of its tomographic images is susceptible to the reconstruction method. In this paper, we investigate the projection x-space reconstruction algorithm for an open-side field-free-line magnetic particle imaging device with partial field-of-view (pFOV), then reconstruct the projection data into tomographic images. In simulation study, we analyze the image in 1D to show the effect of fundamental frequency recovery and in 2D to evaluate the performance of the filtered back-projection algorithm. The overlapping portions of pFOV can be used to recover the offset constant, called the smoothness constraint. 1D profiles are superimposed by pFOVs after fundamental frequency recovery, and the overlapped parts were averaged. Different filters are used to process the image data in filtered back-projection. The results show that the effect of fundamental frequency recovery is significant. The filter in filtered back-projection works similarly to the effect of deconvolution in x-space, with Hann and Hamming filtering being the most effective and least noisy.