US 12,087,433 B2
System and methods for reconstructing medical images using deep neural networks and recursive decimation of measurement data
Hariharan Ravishankar, Bangalore (IN); and Dattesh Dayanand Shanbhag, Bangalore (IN)
Assigned to GE Precision Healthcare LLC, Milwaukee, WI (US)
Filed by GE Precision Healthcare LLC, Wauwatosa, WI (US)
Filed on May 18, 2023, as Appl. No. 18/319,686.
Application 18/319,686 is a continuation of application No. 17/364,544, filed on Jun. 30, 2021, granted, now 11,699,515.
Application 17/364,544 is a continuation of application No. 16/691,430, filed on Nov. 21, 2019, granted, now 11,133,100, issued on Sep. 28, 2021.
Prior Publication US 2023/0290487 A1, Sep. 14, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G16H 30/40 (2018.01); A61B 5/055 (2006.01); A61B 6/03 (2006.01); G06N 3/045 (2023.01); G06T 7/00 (2017.01); G06T 11/00 (2006.01); G16H 30/20 (2018.01)
CPC G16H 30/40 (2018.01) [A61B 5/055 (2013.01); A61B 6/032 (2013.01); G06N 3/045 (2023.01); G06T 7/0014 (2013.01); G06T 11/008 (2013.01); G16H 30/20 (2018.01); G06T 2207/10081 (2013.01); G06T 2207/10084 (2013.01); G06T 2207/10088 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving measurement data acquired by an imaging device;
flattening the measurement data;
selecting a decimation strategy based on a size of the measurement data, wherein decimating the measurement data according to the decimation strategy comprises, sampling the measurement data at a sampling density specified by the decimation strategy, wherein the sampling density is lower than a native sampling density of the measurement data, wherein the measurement data is acquired by an MRI scanner, and the measurement data comprises k-space data of an anatomical region of a patient acquired by the MRI scanner;
producing a reconstructed image from the flattened measurement data using the decimation strategy and a plurality deep neural networks, where the selection is further based on a size of an input layer of the plurality of deep neural networks; and
displaying the reconstructed image via a display device.