Image analysis using deviation from normal data

Abstract

Systems and techniques for facilitating image analysis using deviation from normal data are presented. In one example, a system generates atlas map data indicative of an atlas map that includes a first portion of patient image data from a plurality of reference patients and a second portion of the patient image data from a plurality of target patients. The first portion of the patient image data is matched to a corresponding age group for a set of patient identities associated with the first portion of the patient image data. The system also generates deviation map data that represents an amount of deviation for the second portion of the patient image data compared to the first portion of the patient image data. Furthermore, the system trains a neural network based on the deviation map data to determine one or more clinical conditions.

1 . A 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 first data and second data, wherein the first data satisfies a first defined criterion, and wherein the second data satisfies a second defined criterion; 
 a deviation map component that generates deviation map data that represents an amount of deviation for the second data based on a comparison with the first data, resulting in a deviation map; and 
 a neural network component that uses the deviation map to train a neural network to classify third data as satisfying one of the first defined criterion, the second defined criterion, or a third defined criterion, wherein the first defined criterion, the second defined criterion, and the third defined criterion are different criteria. 
   
     
     
         2 . The system of  claim 1 , wherein the first defined criterion represents a normal condition, wherein the second defined criterion represents an abnormal condition, and wherein the third defined criterion represents a potential abnormal condition. 
     
     
         3 . The system of  claim 1 , wherein the atlas map component formats the atlas map data as a matrix of numerical data values that represent the first data and the second data. 
     
     
         4 . The system of  claim 3 , wherein the deviation map component converts the matrix of numerical data values into a matrix of colorized data values formatted based on the amount of deviation for the second data compared to the first data, wherein a first colorized data value of the matrix of colorized data values corresponds to the first defined criterion, wherein a second colorized data value of the matrix of colorized data values corresponds to the second defined criterion, and wherein a third colorized data value of the matrix of colorized data values corresponds to the third defined criterion. 
     
     
         5 . The system of  claim 1 , wherein the atlas map component normalizes first numerical data values that represent the first data and second numerical data values that represent the second data. 
     
     
         6 . The system of  claim 1 , wherein the atlas map component generates an image intensity value associated with a mean value and a standard deviation value for a data value in the atlas map. 
     
     
         7 . The system of  claim 6 , wherein the deviation map component subtracts the mean value from the image intensity value to generate a difference value, and divides the difference value by the standard deviation value to facilitate generation of the deviation map data. 
     
     
         8 . The system of  claim 1 , wherein the neural network component generates machine learning data to train the neural network. 
     
     
         9 . The system of  claim 8 , wherein the machine learning data comprises a set of filter values for the neural network based on a training phase associated with the deviation map data. 
     
     
         10 . The system of  claim 8 , wherein the machine learning data comprises a set of weights for a set of filters associated with the neural network and based on a training phase associated with the deviation map data. 
     
     
         11 . The system of  claim 1 , wherein the neural network component trains the neural network for generation of a trained neural network. 
     
     
         12 . The system of  claim 1 , where the neural network performs downsampling and upsampling of the deviation map data associated with convolutional layers of the neural network, wherein the downsampling and upsampling is one of sequential downsampling and upsampling, or parallel downsampling and upsampling. 
     
     
         13 . A method, comprising:
 generating, by a system comprising a processor, atlas map data indicative of an atlas map, wherein the atlas map includes first data and second data, wherein the first data satisfies a first defined criterion, and wherein the second data satisfies a second defined criterion;   generating, by the system, a deviation map data that represents an amount of deviation for the second data based on a comparison with the first data, resulting in a deviation map; and   using, by the system, the deviation map to train a neural network to classify third data as satisfying one of the first defined criterion, the second defined criterion, or a third defined criterion, wherein the first defined criterion, the second defined criterion, and the third defined criterion are different criteria.   
     
     
         14 . The method of  claim 13 , further comprising:
 formatting, by the system, the atlas map data as a matrix of numerical data values that represent the first data and the second data.   
     
     
         15 . The method of  claim 14 , further comprising:
 converting, by the system, the matrix of numerical data values into a matrix of colorized data values formatted based on the amount of deviation for the second data compared to the first data, wherein a first colorized data value of the matrix of colorized data values corresponds to the first defined criterion, wherein a second colorized data value of the matrix of colorized data values corresponds to the second defined criterion, and wherein a third colorized data value of the matrix of colorized data values corresponds to the third defined criterion.   
     
     
         16 . The method of  claim 13 , further comprising:
 normalizing, by the system, first numerical data values that represent the first data and second numerical data values that represent the second data.   
     
     
         17 . The method of  claim 13 , wherein the first defined criterion represents a normal condition, wherein the second defined criterion represents an abnormal condition, and wherein the third defined criterion represents a potential abnormal condition. 
     
     
         18 . A non-transitory computer readable storage device comprising instructions that, in response to execution, cause a system comprising a processor to perform operations, comprising:
 generating atlas map data indicative of an atlas map, wherein the atlas map includes first data and second data, wherein the first data satisfies a first defined criterion, and wherein the second data satisfies a second defined criterion;   generating deviation map data that indicates an amount of deviation for the second data based on a comparison with the first data, resulting in a deviation map; and   training a neural network to classify third data as satisfying one of the first defined criterion, the second defined criterion, or a third defined criterion, wherein the training comprises using the deviation map to train the neural network, wherein the first defined criterion, the second defined criterion, and the third defined criterion are different criteria.   
     
     
         19 . The non-transitory computer readable storage device of  claim 18 , wherein the operations further comprise:
 generating an image intensity value associated with a mean value and a standard deviation value for a data value in the atlas map.   
     
     
         20 . The non-transitory computer readable storage device of  claim 19 , wherein the operations further comprise:
 subtracting the mean value from the image intensity value to generate a difference value; and   dividing the difference value by the standard deviation value to facilitate generation of the deviation map data.