US 12,080,021 B2
Training a machine learning algorithm using digitally reconstructed radiographs
Kim Le, Munich (DE); Yannic Meurer, Laim (DE); Thomas Drexl, Poing (DE); and Thomas Feilkas, Kirchseeon (DE)
Assigned to BRAINLAB AG, Munich (DE)
Appl. No. 17/274,580
Filed by Brainlab AG, Munich (DE)
PCT Filed Sep. 20, 2019, PCT No. PCT/EP2019/000277
§ 371(c)(1), (2) Date Mar. 9, 2021,
PCT Pub. No. WO2021/052552, PCT Pub. Date Mar. 25, 2021.
Prior Publication US 2021/0383565 A1, Dec. 9, 2021
Int. Cl. G06T 7/73 (2017.01); G06N 3/04 (2023.01); G06N 3/08 (2023.01); G06T 7/00 (2017.01); G16H 30/20 (2018.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01)
CPC G06T 7/73 (2017.01) [G06N 3/04 (2013.01); G06N 3/08 (2013.01); G06T 7/0012 (2013.01); G16H 30/20 (2018.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01); G06T 2207/10121 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30004 (2013.01)] 13 Claims
OG exemplary drawing
 
1. A computer-implemented method of training a machine learning model for determining position of an image representation of an annotated anatomical structure in a two-dimensional x-ray image, the method comprising the following steps:
acquiring atlas data which describes a three-dimensional shape of the anatomical structure and at least one projection parameter for generating two-dimensional x-ray images of the anatomical structure;
acquiring image training data which describes the two-dimensional x-ray images being digitally reconstructed radiographs and including an image representation of the anatomical structure, wherein the two-dimensional x-ray images are generated based on the at least one projection parameter that is described in the atlas data for the anatomical structure;
acquiring annotation data which describes an annotation for the anatomical structure; and
determining model parameter data which describes model parameters of the machine learning model for establishing a relation between the anatomical structure in the two-dimensional x-ray images and the annotation,
wherein the model parameter data is determined by training the model parameter data through inputting the image training data and the annotation data into a function which establishes the relation, the function being part of the machine learning model.