US 11,941,732 B2
Multi-slice MRI data processing using deep learning techniques
Xiao Chen, Lexington, MA (US); Zhang Chen, Brookline, MA (US); Shanhui Sun, Lexington, MA (US); and Terrence Chen, Lexington, MA (US)
Assigned to Shanghai United Imaging Intelligence Co., Ltd., Shanghai (CN)
Filed by Shanghai United Imaging Intelligence Co., Ltd., Shanghai (CN)
Filed on Oct. 28, 2021, as Appl. No. 17/513,320.
Prior Publication US 2023/0135995 A1, May 4, 2023
Int. Cl. G06T 11/00 (2006.01); A61B 5/00 (2006.01); A61B 5/055 (2006.01); G06T 7/00 (2017.01)
CPC G06T 11/008 (2013.01) [A61B 5/055 (2013.01); A61B 5/7267 (2013.01); G06T 7/0012 (2013.01); G06T 2207/10088 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30016 (2013.01); G06T 2207/30048 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for reconstructing magnetic resonance imaging (MRI) images, the method comprising:
obtaining a simultaneous multi-slice (SMS) dataset, wherein the SMS dataset comprises first under-sampled MRI data associated with a first MRI slice of an organ and second under-sampled MRI data associated with a second MRI slice of the organ, the first MRI slice and the second MRI slice simultaneously acquired from an MRI procedure; and
generating, using an artificial neural network (ANN), a first reconstructed MRI image of the organ corresponding to the first MRI slice and a second reconstructed MRI image of the organ corresponding to the second MRI slice, wherein the ANN is trained for generating the first reconstructed MRI image and the second reconstructed MRI image, and the training comprises:
processing first under-sampled MRI training data of an SMS training dataset through an instance of the ANN to obtain a first estimated MRI image, the first under-sampled MRI training data corresponding to a first MRI slice of the SMS training dataset;
processing second under-sampled MRI training data of the SMS training dataset through the instance of the ANN to obtain a second estimated MRI image, the second under-sampled MRI training data corresponding to a second MRI slice of the SMS training dataset;
determining a combined training loss by jointly considering a first training loss associated with the first estimated MRI image and a second training loss associated with the second estimated MRI image; and
adjusting parameters of the instance of the ANN based on a gradient descent of the combined training loss.