US 12,249,008 B2
Implicit neural representation learning with prior embedding for sparsely sampled image reconstruction and other inverse problems
Liyue Shen, Cambridge, MA (US); and Lei Xing, Palo Alto, CA (US)
Assigned to The Board of Trustees of the Leland Stanford Junior University, Stanford, CA (US)
Filed by The Board of Trustees of the Leland Stanford Junior University, Stanford, CA (US)
Filed on Sep. 14, 2022, as Appl. No. 17/944,709.
Application 17/944,709 is a continuation in part of application No. 17/835,896, filed on Jun. 8, 2022, abandoned.
Claims priority of provisional application 63/210,433, filed on Jun. 14, 2021.
Prior Publication US 2023/0024401 A1, Jan. 26, 2023
Int. Cl. G06N 3/08 (2023.01); G06T 11/00 (2006.01)
CPC G06T 11/006 (2013.01) [G06N 3/08 (2013.01); G06T 11/005 (2013.01); G06T 2210/41 (2013.01)] 6 Claims
OG exemplary drawing
 
1. A method for diagnostic imaging reconstruction comprising:
storing a prior image xpr from a scan of a subject, comprising image intensity at each coordinate in image space;
initializing parameters of a neural network using the prior image xpr;
wherein the neural network maps coordinates in image space to corresponding intensity values in the prior image;
wherein initializing the parameters comprises minimizing an objective function representing a difference between intensity values of the prior image and predicted intensity values output from the neural network, thereby creating an implicit neural representation of the prior image;
performing a scan to acquire subsampled (sparse) measurements y of the subject;
training the neural network using the measurements y to learn a neural representation of a reconstructed image x, wherein the training comprises minimizing an objective function representing a difference between the measurements y and a forward model applied to predicted image intensity values output from the neural network;
computing image intensity values output from the trained neural network from coordinates in image space input to the trained neural network to produce predicted image intensity values.