Image analysis using deviation from normal data

  • US 11,093,820 B2
  • Filed: 12/27/2017
  • Issued: 08/17/2021
  • Est. Priority Date: 10/19/2017
  • Status: Active Grant
First Claim
Patent Images

1. A machine learning system, comprising:

  • a memory that stores computer executable components; and

    a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise;

    an atlas map component that generates atlas map data indicative of an atlas map that includes a set of normal patient image data for a plurality of first pediatric patients and a set of abnormal patient image data for a plurality of second pediatric patients, different from the plurality of first pediatric patients, wherein the set of normal patient image data and the set of abnormal patient image data are respective pediatric brain scan images, and wherein normal patient image data of the set of normal patient image data and abnormal patient image data of the set of abnormal patient image data is assigned to respective pediatric age groups based on respective patient identities;

    a deviation map component that converts numerical data values included in the atlas map data into data values formatted based on a first amount of deviation of the set of abnormal patient image data compared to the set of normal patient image data and generates, based on the data values, deviation map data that represents a second amount of deviation for the set of abnormal patient image data compared to the normal patient image data; and

    a neural network component that trains a neural network based on the deviation map data to determine respective clinical conditions for another set of patient image data of a plurality of third pediatric patients, different from the plurality of first pediatric patients and the plurality of second pediatric patients.

View all claims
    ×
    ×

    Thank you for your feedback

    ×
    ×