Deep convolutional neural network with self-transfer learning

Abstract

Systems and techniques for facilitating a deep convolutional neural network with self-transfer learning are presented. In one example, a system includes a machine learning component, a medical imaging diagnosis component and a visualization component. The machine learning component generates learned medical imaging output regarding an anatomical region based on a convolutional neural network that receives medical imaging data. The machine learning component also performs a plurality of sequential downsampling and upsampling of the medical imaging data associated with convolutional layers of the convolutional neural network. The medical imaging diagnosis component determines a classification and an associated localization for a portion of the anatomical region based on the learned medical imaging output associated with the convolutional neural network. The visualization component generates a multi-dimensional visualization associated with the classification and the localization for the portion of the anatomical region.

What is claimed is: 
     
         1 . A convolutional neural network system, comprising:
 a processor; and   a memory that stores executable instructions that, when executed by the processor, facilitate performance of operations, comprising:
 performing a first convolutional layer process associated with sequential downsampling of imaging data followed by a second convolutional layer process associated with sequential upsampling of the imaging data to facilitate generation of a learned imaging output, wherein a first convolutional layer of the first convolutional layer process corresponds to a last convolutional layer of the second convolutional layer process; and 
 predicting a condition associated with the learned imaging output. 
   
     
     
         2 . The convolutional neural network system of  claim 1 , wherein the operations further comprise:
 employing a convolutional neural network associated with the first convolutional layer process and the second convolutional layer process to generate the learned imaging output.   
     
     
         3 . The convolutional neural network system of  claim 1 , wherein performing the first convolutional layer process comprises performing the first convolutional layer process based on a first convolutional layer filter that comprises a first size and a second convolutional layer filter that comprises a second size that is different than the first size, and wherein performing the second convolutional layer process comprises performing the second convolutional layer process based on the first convolutional layer filter that comprises the first size and the second convolutional layer filter that comprises the second size. 
     
     
         4 . The convolutional neural network system of  claim 1 , wherein the operations further comprise:
 performing the sequential upsampling of the imaging data in a reverse sampling sequence with respect to the sequential downsampling of the imaging data.   
     
     
         5 . The convolutional neural network system of  claim 1 , wherein the generation of the learned imaging output is based on a class activation mapping process that applies a set of weights to a set of heat maps associated with the imaging data. 
     
     
         6 . The convolutional neural network system of  claim 1 , wherein the operations further comprise:
 determining a classification for the condition associated with the learned imaging output.   
     
     
         7 . The convolutional neural network system of  claim 1 , wherein the operations further comprise:
 determining a localization for at least a portion of the learned imaging output.   
     
     
         8 . The convolutional neural network system of  claim 1 , wherein the operations further comprise:
 generating a multi-dimensional visualization related to the condition.   
     
     
         9 . A method, comprising:
 performing, by a system comprising a processor, training a convolutional neural network, using medical imaging data, to generate a learned medical imaging output regarding a patient body; and   predicting, by the system, a medical condition associated with a portion of the patient body based on the learned medical imaging output.   
     
     
         10 . The method of  claim 9 , further comprising:
 analyzing the medical imaging data based on a first filter that comprises a first size;   analyzing the medical imaging data based on a second filter that comprises a second size that is different than the first size; and   analyzing the medical imaging data based on a third filter that comprises the first size associated with the first filter.   
     
     
         11 . The method of  claim 9 , further comprising:
 generating the learned medical imaging output based on a first convolutional layer process associated with the downsampling of the medical imaging data and a second convolutional layer process associated with the upsampling of the medical imaging data.   
     
     
         12 . The method of  claim 9 , further comprising:
 performing the upsampling of the medical imaging data in a reverse sampling sequence with respect to the downsampling of the medical imaging data.   
     
     
         13 . The method of  claim 9 , wherein the predicting the medical condition associated with the portion of the patient body comprises determining a classification for the medical condition associated with the portion of the patient body based on the learned medical imaging output. 
     
     
         14 . The method of  claim 9 , wherein the predicting the medical condition associated with the portion of the patient body comprises determining a localization for the portion of the patient body associated with the medical condition based on the learned medical imaging output. 
     
     
         15 . The method of  claim 12 , further comprising:
 generating, by the system, a multi-dimensional visualization related to the medical condition associated with the portion of the patient body.   
     
     
         16 . A computer readable storage device comprising instructions that, in response to execution, cause a system comprising a processor to perform operations, comprising:
 training a convolutional neural network, using medical imaging data, to generate a learned medical imaging output regarding an anatomical region; and   predicting a medical condition associated with an anatomical region based on learned medical imaging output.   
     
     
         17 . The computer readable storage device of  claim 16 , wherein the training of the convolutional neural network comprises employing a first portion of the medical imaging data for training associated with the convolutional neural network, and wherein the operations further comprise:
 employing a second portion of the medical imaging data for validation associated with the convolutional neural network; and   employing a third portion of the medical imaging data for testing associated with the convolutional neural network.   
     
     
         18 . The computer readable storage device of  claim 16 , wherein the operations further comprise:
 modifying an orientation of a set of medical images for data augmentation of the medical imaging data.   
     
     
         19 . The computer readable storage device of  claim 18 , wherein the modifying of the orientation of the set of medical images comprises modifying of the orientation of the set of medical images through an affine transformation for the data augmentation associated with the medical imaging data. 
     
     
         20 . The computer readable storage device of  claim 16 , wherein the operations further comprise:
 determining a location of a disease associated with the medical condition.