An Apparatus and Process For Optimizing Radiation Detection Counting Times Using Machine Learning


Technology Description


This algorithm is applicable to radiation detection and counting, for example, in Gamma spectroscopy. The purpose of this algorithm is to eliminate the need for a priori information when counting radioactive material and instead using machine learning techniques to solve the optimal counting time real-time. Data is arranged into a matrix, which can then be used to create training material for classification. Once the desired outcome is achieved (i.e., radiation source is identified or radiation dose reaches a specific threshold), the radiation source is turned off.


Features & Benefits


  • Multiple parallel machine learning methods achieve optimal counting time
  • User defines confidence level/desired resolution, rather than count time
  • Multiple sensing channels are used to identify radiation




  • Passive and active interrogation systems located at border crossings
  • Common gamma spectroscopy systems
  • Diagnostic imaging applications


Background of Invention


There are a number of systems that exist which are used to detect radiation or determine the content of radioactive material. They are applicable across various nuclear science applications. Most of the current solutions on the market require a radiation detector, process signal electronics and an analytical approach to determining the amount of radioactive material existing. This approach typically involves collecting detector data over a finite period of time and post processing. This algorithm outlines a method in which a counting architecture called Frieder Counting is used to collect raw count data from a counting system. Then machine learning, data mining and data analytics techniques are used to optimize the counting time of radiation or radionuclides present in a material. The advantage of this solution is that it can be applied in real time whereas the current solutions cannot.




Patent pending; seeking development partners





Patent Information:
Tech ID:
David Dickson
IP & Licensing Manager
Oregon State University
Steve Reese
Jessica Curtis
Ophir Frieder
© 2018. All Rights Reserved. Powered by Inteum