US 11,890,124 B2
Systems and methods for low-dose AI-based imaging
Andre D. A. Souza, Boylston, MA (US); Michael Philip Marrama, Ayer, MA (US); Patrick A. Helm, Milton, MA (US); Mehdi Rahman, North Billerica, MA (US); Kyo C. Jin, Durham, NH (US); and Michael D. Ketcha, Baltimore, MD (US)
Assigned to Medtronic Navigation, Inc., Louisville, CO (US)
Filed by Medtronic Navigation, Inc., Louisville, CO (US)
Filed on Feb. 1, 2021, as Appl. No. 17/163,719.
Prior Publication US 2022/0240879 A1, Aug. 4, 2022
Int. Cl. A61B 6/00 (2006.01); G06T 11/00 (2006.01); G16H 50/20 (2018.01); G06N 3/08 (2023.01)
CPC A61B 6/5205 (2013.01) [A61B 6/5258 (2013.01); G06N 3/08 (2013.01); G06T 11/005 (2013.01); G06T 11/006 (2013.01); G06T 11/008 (2013.01); G16H 50/20 (2018.01); A61B 6/4458 (2013.01); A61B 6/547 (2013.01); A61B 6/563 (2013.01); G06T 2210/41 (2013.01); G06T 2211/421 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A low-dose imaging method, comprising:
receiving a sparse image set of a portion of a patient's anatomy, wherein the sparse image set comprises a plurality of globally-distributed artifacts;
up-sampling the sparse image set, in a sinogram domain and using a first neural network trained in the sinogram domain, to yield an up-sampled sinogram in which the first neural network has filled in at least some of the plurality of globally-distributed artifacts;
generating, from the up-sampled sinogram, an initial reconstruction; and
removing, from the initial reconstruction and using a second neural network trained in a slice domain, one or more artifacts in the initial reconstruction to yield a final output volume.