US 11,714,152 B2
Methods for scan-specific artifact reduction in accelerated magnetic resonance imaging using residual machine learning algorithms
Mehmet Akcakaya, Minneapolis, MN (US); Steen Moeller, Minneapolis, MN (US); and Chi Zhang, Minneapolis, MN (US)
Assigned to REGENTS OF THE UNIVERSITY OF MINNESOTA, Minneapolis, MN (US)
Filed by REGENTS OF THE UNIVERSITY OF MINNESOTA, Minneapolis, MN (US)
Filed on Apr. 27, 2020, as Appl. No. 16/858,922.
Claims priority of provisional application 62/839,073, filed on Apr. 26, 2019.
Prior Publication US 2020/0341103 A1, Oct. 29, 2020
Int. Cl. G01R 33/58 (2006.01); G01R 33/56 (2006.01); G06N 3/08 (2023.01); G06N 3/048 (2023.01); G01R 33/565 (2006.01)
CPC G01R 33/58 (2013.01) [G01R 33/5608 (2013.01); G06N 3/048 (2023.01); G06N 3/08 (2013.01); G01R 33/56545 (2013.01)] 21 Claims
OG exemplary drawing
 
1. A method for reconstructing an image from undersampled k-space data acquired with a magnetic resonance imaging (MRI) system, the steps of the method comprising:
(a) providing to a computer system, undersampled k-space data and calibration data acquired with an MRI system;
(b) processing the calibration data with a computer system to learn parameters for a residual machine learning algorithm implemented with a hardware processor and memory of the computer system, wherein the residual machine learning algorithm comprises a linear residual connection and a multi-layer portion having at least one nonlinear processing layer;
(c) estimating missing k-space data by inputting the undersampled k-space data to the residual machine learning algorithm, generating output as the missing k-space data; and
(d) reconstructing an image from the undersampled k-space data and the estimated missing k-space data.