Skip to main content
Science and Technology Partnerships and Outreach

Machine Learning Technique to Identify Grains in Polycrystalline Materials Samples (ANL-IN-16-126)

The invention comprises a method to automate the process of identifying grains in a polycrystalline material, and hence can process large volumes of data with enhanced accuracy.
Intellectual Property Available to License
US Patent 10,839,195 B2
  • Machine Learning Technique to Identify Grains in Polycrystalline Materials Samples (ANL-IN-16-126)

The invention comprises a method to automate the process of identifying grains in a polycrystalline material, and hence can process large volumes of data with enhanced accuracy.

Applications:

Such a technique is vital for the real time analysis of data from large characterization facilities such as synchrotrons but is also applicable to any 3D crystallographic data of polycrystalline samples. The algorithm  provides a method to quickly identify, track faults and quantify features that affect material properties. This is potentially of interest to:

  • Defense industry (e.g., armor integrity analysis)
  • Infrastructure (e.g., steel and other metal processing companies)
  • Electronics industry (e.g., thin film microstructure)
  • Manufacturing companies (e.g., automotive, aerospace, etc.)