Labeling training data using large amounts of navigation data

Abstract

Disclosed herein are methods and systems for automatically labeling image data for machine learning training purposes. The method includes retrieving a set of navigation data and image data from a set of egos navigating through an environment including at least one feature, generating a three-dimensional (3D) model of the environment using the navigation data and image data of at least a subset of the set of egos, the 3D model including a virtual representation of the at least one feature of the environment, identifying a machine learning label associated with the at least one feature in the image data, receiving second navigation data and second image data from a second ego not included in the set of egos, the second ego navigating through the environment and the second image data including the at least one feature, and automatically generating a machine learning label for the at least one feature depicted in the second image data.

Images