Patent · US Active

Generation of synthetic images for training a neural network model

US10867214B2 · kind B2 · utility

4Cited by
9References
20Claims
0Family size

Assignee

Inventors

Key dates

Filing dateJan 24, 2019
Grant dateDec 15, 2020
Priority date
Expiry dateFeb 8, 2039

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06V2201/07
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

Training deep neural networks requires a large amount of labeled training data. Conventionally, labeled training data is generated by gathering real images that are manually labelled which is very time-consuming. Instead of manually labelling a training dataset, domain randomization technique is used generate training data that is automatically labeled. The generated training data may be used to train neural networks for object detection and segmentation (labelling) tasks. In an embodiment, the generated training data includes synthetic input images generated by rendering three-dimensional (3D) objects of interest in a 3D scene. In an embodiment, the generated training data includes synthetic input images generated by rendering 3D objects of interest on a 2D background image. The 3D objects of interest are objects that a neural network is trained to detect and/or label.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.