US 2019/0377980 A1 · GENERATING TRAINING DATA FOR MACHINE LEARNING CLASSIFIER

Abstract

Data representing a scene is received. The scene includes labeled elements such as walls, a floor, a ceiling, and objects placed at various locations in the scene. The original received scene may be modified in different ways to create new scenes that are based on the original scene. These modifications include adding clutter to the scene, moving one or more elements of the scene, swapping one or more elements of the scene with different labeled elements, changing the size, color, or materials associated with one or more of the elements of the scene, and changing the lighting used in the scene. Each new scene may be used to generate labeled training data for a classifier by placing a virtual sensor (e.g., a camera) in the new scene, and generating sensor output data for the virtual sensor based on its placement in the new scene.

Claims20

Dependent claims

A system for generating training data for machine learning classifiers, the system comprising:

at least one computing device;
and
a training engine that:
receives a first scene, wherein the first scene includes a plurality of elements, and each element is associated with a label;
generates a second scene from the first scene by performing one or more of:
changing a lighting characteristic of the first scene;
adding one more elements to the plurality of elements of the first scene;
and
changing a characteristic of one or more elements of the plurality of elements of the first scene;
generates training data from the second scene;
and
trains a machine learning classifier using the generated training data.

9 more elements

Dependencies are inferred from the claim text.

Generating training data for machine learning classifier

Life of this patent

Generating training data for machine learning classifier

Life of this patent

Abstract

Claims20