Patent · US Active

Systems and methods for optical material characterization of waste materials using machine learning

US11069053B2 · kind B2 · utility

4Cited by
4References
33Claims
0Family size

Assignee

Inventors

Key dates

Filing dateOct 31, 2018
Grant dateJul 20, 2021
Priority date
Expiry dateMay 28, 2039

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06V20/52
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

Systems and methods for optical material characterization of waste materials using machine learning are provided. In one embodiment, a system comprises: an imaging device configured to generate image frames an area and target objects within the area; an object characterization processor coupled to the imaging device and comprising Neural Processing Units and a Neural Network Parameter Set. The Neural Network Parameter Set stores learned parameters utilized by the one or more Neural Processing Units for characterizing the one or more target objects. The Neural Processing Units are configured by the Neural Network Parameter Set to detect a presence of a plurality of different materials within the image frames based on a plurality of different features. For a first image frame of the plurality of image frames, the Neural Processing Units outputs material characterization data that identifies which of the plurality of different materials are detected in the first image frame.

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