US 11,854,126 B2
Methods and apparatus for deep learning based image attenuation correction
Joshua Schaefferkoetter, Knoxville, TN (US)
Assigned to Siemens Medical Solutions USA, Inc., Malvern, PA (US)
Filed by Siemens Medical Solutions USA, Inc., Malvern, PA (US)
Filed on Jul. 7, 2021, as Appl. No. 17/305,395.
Prior Publication US 2023/0009528 A1, Jan. 12, 2023
Int. Cl. G06T 11/00 (2006.01); G06T 7/00 (2017.01); G06T 3/00 (2006.01); G06N 3/08 (2023.01)
CPC G06T 11/008 (2013.01) [G06N 3/08 (2013.01); G06T 3/0068 (2013.01); G06T 7/0012 (2013.01); G06T 2207/10104 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receiving positron emission tomography (PET) measurement data and anatomy measurement data from an image scanning system;
generating PET images based on the PET measurement data and anatomy images based on the anatomy measurement data;
training a neural network with the PET images and the anatomy images, the neural network comprising a transformation stage and a registration stage, wherein the training configures the transformation stage to generate initial attenuation maps based on the anatomy measurement data, and the training configures the registration stage to generate final attenuation maps based on the initial attenuation maps and the PET images;
generating attenuation maps based on the training; and
storing the trained neural network in a data repository.