Machine learning training corpus apparatus and method
AssigneeSiemens Healthineers International AG
InventorsAngelo Genghi, Mário João Fartaria, Tomasz Morgas, Thomas Coradi, Pascal Paysan, Adam Michal Strzelecki
Life of this patent
Estimating the term…
Abstract
A control circuit accesses a plurality of computed tomography (CT) information items and generates a plurality of synthetic cone-beam computed tomography (CBCT) information items as a function thereof. These teachings can then provide for generating a machine learning training corpus as a function of paired data comprising pairs of the synthetic CBCT information items with other information items. Those other information items may comprise, for example, one or more of the aforementioned CT information items and/or structure information.
Claims19
A method of generating a machine learning training corpus comprising the steps of:
- accessing a plurality of computed tomography (CT) information items;
- generating a plurality of synthetic cone-beam computed tomography (CBCT) information items as a function of the plurality of computed tomography information items;
- generating the machine learning training corpus as a function of paired data comprising pairs of the synthetic CBCT information items with other information items comprising at least one of target volumes and organs at risk, such that a machine learning model trained with the machine learning training corpus will learn a mapping between the synthetic CBCT information items and the other information items.
An apparatus configured to generate a machine learning training corpus comprising:
- a control circuit configured to:
- access a plurality of computed tomography (CT) images;
- generate a plurality of synthetic cone-beam computed tomography (CBCT) images as a function of the plurality of computed tomography images;
- and
- generate the machine learning training corpus as a function of paired imaging data comprising pairs of the CT images with the synthetic CBCT images, such that a machine learning model trained with the machine learning training corpus will learn a mapping between the synthetic CBCT images and the CT images.
A method comprising:
- by a control circuit:
- accessing a machine learning model that has been trained using a training corpus that includes paired imaging data comprising pairs of computed tomography (CT) images with synthetic cone-beam computed tomography (CBCT) images, such that the machine learning model learned a mapping between the synthetic CBCT information items and the CT images;
- automatically processing patient image information using the machine learning model to generate corresponding output.
Dependencies are inferred from the claim text.