US 12,456,166 B2
Magnetic particle imaging reconstruction method based on neural network constrained by forward model
Xueli Chen, Xi'an (CN); Shenghan Ren, Xi'an (CN); Pengfei Huang, Xi'an (CN); Duofang Chen, Xi'an (CN); Hui Xie, Xi'an (CN); Shouping Zhu, Xi'an (CN); and Jie Tian, Xi'an (CN)
Assigned to Xidian University, Xi'an (CN)
Filed by Xidian University, Xi'an (CN)
Filed on May 31, 2023, as Appl. No. 18/326,043.
Claims priority of application No. 202211002854.4 (CN), filed on Aug. 19, 2022.
Prior Publication US 2024/0062337 A1, Feb. 22, 2024
Int. Cl. G06T 5/10 (2006.01); G01R 33/12 (2006.01); G06T 5/70 (2024.01)
CPC G06T 5/10 (2013.01) [G01R 33/1276 (2013.01); G06T 5/70 (2024.01); G06T 2207/20084 (2013.01)] 8 Claims
OG exemplary drawing
 
1. A magnetic particle imaging reconstruction method based on a neural network constrained by a forward model, comprising:
step 1: obtaining measured data of magnetic particle imaging and data of a system matrix obtained through calibration, then performing Fourier transform to obtain frequency-domain data, and performing frequency feature screening by using a signal-to-noise ratio threshold;
step 2: building a reconstruction network by using a Pytorch, to realize mapping from one-dimensional voltage data to a multi-dimensional magnetic particle concentration distribution;
step 3: using the system matrix as the forward model of magnetic particle imaging, and generating simulated voltage data by a reconstructed multi-dimensional magnetic particle concentration distribution; and
step 4: obtaining a loss function according to a difference between the simulated voltage data and input voltage data, adding a total variation norm of concentration, and updating parameters of the reconstruction network.