Systems and methods for deep model translation generation

Abstract

Embodiments of the present invention relate to systems and methods for improving the training of machine learning systems to recognize certain objects within a given image by supplementing an existing sparse set of real-world training images with a comparatively dense set of realistic training images. Embodiments may create such a dense set of realistic training images by training a machine learning translator with a convolutional autoencoder to translate a dense set of synthetic images of an object into more realistic training images. Embodiments may also create a dense set of realistic training images by training a generative adversarial network (“GAN”) to create realistic training images from a combination of the existing sparse set of real-world training images and either Gaussian noise, translated images, or synthetic images. The created dense set of realistic training images may then be used to more effectively train a machine learning object recognizer to recognize a target object in a newly presented digital image.

The invention claimed is: 
     
         1 . A method for improving the training of a computer-based object recognizer to recognize a target object within an image selected from a set of real-world images, the method comprising:
 training a generative adversarial network (GAN) to produce a plurality of generated images of the target object, said GAN comprising a computer-based machine learning system;   training the computer-based object recognizer to recognize the target object within a newly presented digital image by providing the object recognizer a collection of training images, said collection of training images assembled from:
 (1) a set of real-world images of the target object, and 
 (2) the plurality of generated images of the target object. 
   
     
     
         2 . The method of  claim 1  wherein the generated images are produced by providing the GAN with Gaussian noise. 
     
     
         3 . The method of  claim 1  wherein the computer-based machine learning system has a discriminative neural network instantiated with the set of real-world images. 
     
     
         4 . The method of  claim 1  wherein the computer-based machine learning system has a discriminative neural network instantiated with the set of real-world images, where each image in the set of real-world images is labeled according to the target object. 
     
     
         5 . The method of  claim 1  wherein the computer-based machine learning system has a generative neural network instantiated with Gaussian noise. 
     
     
         6 . The method of  claim 1  wherein the computer-based machine learning system has a discriminative neural network and a generative neural network, and wherein GAN training includes instantiating the generative neural network with Gaussian noise and instantiating the discriminative neural network with the set of real-world images, where each image in the set of real-world images is labeled according to the target object. 
     
     
         7 . The method of  claim 3  further comprising:
 creating a set of synthetic images of the target object; 
 wherein the GAN training includes instantiating the discriminative neural network with the set of synthetic images. 
 
     
     
         8 . The method of  claim 3  further comprising:
 obtaining a set of translated images; 
 wherein the GAN training includes instantiating the discriminative neural network with the set of translated images. 
 
     
     
         9 . The method of  claim 1  where the object recognizer includes a loss function that is invoked iteratively during the object recognizer training. 
     
     
         10 . A computer-based object recognizer for recognizing a target object within an image selected from a set of real-world images, comprising:
 a trainable machine learning processor;   a generative adversarial network (GAN) configured to produce a plurality of generated images of the target object;   a training processor programmed to supply to the trainable machine learning processor with a collection of training images, assembled from:
 (1) a set of real-world images of the target object, and 
 (2) the plurality of generated images of the target object. 
   
     
     
         11 . The computer-based object recognizer of  claim 10  further comprising a source of Gaussian noise, and wherein the GAN is configured ingest the Gaussian noise as an input used to produce the plurality of generated images of the target object. 
     
     
         12 . The computer-based object recognizer of  claim 10  wherein the GAN includes a discriminative neural network instantiated with the set of real-world images. 
     
     
         13 . The computer-based object recognizer of  claim 10  wherein the GAN includes a discriminative neural network instantiated with the set of real-world images, where each image in the set of real-world images is labeled according to the target object. 
     
     
         14 . The computer-based object recognizer of  claim 10  wherein the GAN includes a generative neural network instantiated with Gaussian noise. 
     
     
         15 . The computer-based object recognizer of  claim 10  wherein the GAN includes a discriminative neural network and a generative neural network, and wherein GAN is trained by instantiating the generative neural network with Gaussian noise and by instantiating the discriminative neural network with the set of real-world images, where each image in the set of real-world images is labeled according to the target object. 
     
     
         16 . The computer-based object recognizer of  claim 12  further comprising:
 a set of synthetic images of the target object stored in a computer-readable medium; 
 wherein the GAN is trained by instantiating the discriminative neural network with the set of synthetic images. 
 
     
     
         17 . The computer-based object recognizer of  claim 12  further comprising:
 a set of synthetic images of the target object stored in a computer-readable medium; 
 wherein the GAN is trained by instantiating the discriminative neural network with the set of translated images. 
 
     
     
         18 . The computer-based object recognizer of  claim 10  where the object recognizer includes a loss function that is invoked iteratively during the training of the trainable machine learning processor.