US 11,756,189 B2
Method and systems for providing synthetic labeled training datasets and applications thereof
Simon Hergenröder, Erlangen (DE); Maximilian Metzner, Bavaria (DE); and Sören Weissert, Aalen (DE)
Assigned to Siemens Aktiengesellschaft, Munich (DE)
Appl. No. 17/927,839
Filed by Siemens Aktiengesellschaft, Munich (DE)
PCT Filed May 14, 2021, PCT No. PCT/EP2021/062809
§ 371(c)(1), (2) Date Nov. 25, 2022,
PCT Pub. No. WO2021/239476, PCT Pub. Date Dec. 2, 2021.
Claims priority of application No. 20176860 (EP), filed on May 27, 2020.
Prior Publication US 2023/0196734 A1, Jun. 22, 2023
Int. Cl. G06K 9/00 (2022.01); G06V 10/774 (2022.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01); G06T 7/73 (2017.01); G06T 7/00 (2017.01)
CPC G06V 10/774 (2022.01) [G06T 7/001 (2013.01); G06T 7/74 (2017.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30141 (2013.01)] 15 Claims
OG exemplary drawing
 
1. A computer-implemented method for providing a labeled training dataset, the method comprising:
selecting at least one electronic component in a CAD model of an electronic assembly, the electronic assembly comprising a plurality of electronic components, wherein the CAD model contains a description of the at least one electronic component and a coordinate of the at least one electronic component, wherein the description of the at least one electronic component includes information about a type of the at least one electronic component, and the coordinates of the electronic components are used to determine their positions in the electronic assembly;
generating a plurality of different render images, wherein the different render images include the at least one selected electronic component, wherein a position of the at least one selected electronic component differs in a case of at least two different render images;
labeling the different render images based on the description of the at least one electronic component contained in the CAD model and on the coordinate of the at least one electronic component contained in the CAD model; and
providing a labeled training dataset based on the labeled render images, wherein a label of the at least one electronic component has a description of the at least one electronic component and a position of the at least one electronic component in the render image.