US 11,940,519 B2
Method and system for determining a magnetic susceptibility distribution
Kieran O'Brien, West End (AU); Jin Jin, Chapel Hill (AU); Steffen Bollmann, St Lucia (AU); Markus Barth, St Lucia (AU); and Francesco Cognolato, Casalserugo Pd (IT)
Assigned to Siemens Healthineers AG, Erlangen (DE)
Filed by Siemens Healthcare GmbH, Erlangen (DE); and The University of Queensland, Brisbane (AU)
Filed on Apr. 21, 2022, as Appl. No. 17/725,610.
Claims priority of provisional application 63/177,719, filed on Apr. 21, 2021.
Prior Publication US 2022/0342022 A1, Oct. 27, 2022
Int. Cl. G01R 33/56 (2006.01); G06N 3/045 (2023.01); G06N 3/084 (2023.01); G06V 10/82 (2022.01)
CPC G01R 33/5608 (2013.01) [G06N 3/045 (2023.01); G06N 3/084 (2013.01); G06V 10/82 (2022.01)] 15 Claims
OG exemplary drawing
 
1. A training method for training a first artificial neural network and a second artificial neural network to determine a magnetic susceptibility distribution of a sample, the training method comprising:
storing a simulated magnetic susceptibility map in a computer-readable fashion;
combining an influence of one or more external magnetic susceptibility sources with the simulated magnetic susceptibility map to generate a modified magnetic susceptibility map;
storing the modified magnetic susceptibility map in a computer-readable fashion;
applying a predetermined quantitative susceptibility mapping model to the modified magnetic susceptibility map to generate a first training image;
storing the first training image in a computer-readable fashion;
applying the first artificial neural network to the first training image and applying the second artificial neural network to an output of the first artificial neural network; and
changing first network parameters of the first artificial neural network and second network parameters of the second artificial neural network based on a deviation of an output of the second artificial neural network from the simulated magnetic susceptibility map.