Deep neural network based identification of realistic synthetic images generated using a generative adversarial network
Techniques are provided for deep neural network (DNN) identification of realistic synthetic images generated using a generative adversarial network (GAN). According to an embodiment, a system is described that can comprise a memory that stores computer executable components and a processor that executes the computer executable components stored in the memory. The computer executable components can comprise, a first extraction component that extracts a subset of synthetic images classified as non-real like as opposed to real-like, wherein the subset of synthetic images were generated using a GAN model. The computer executable components can further comprise a training component that employs the subset of synthetic images and real images to train a DNN network model to classify synthetic images generated using the GAN model as either real-like or non-real like.
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:
a training component that trains a deep neural network model to classify synthetic images generated using a synthetic image generation model as either real-like or non-real like using a training dataset comprising first real images and first synthetic images generated using the synthetic image generation model and classified as non-real like, resulting in a trained deep neural network model;
an inference component that employs the trained deep neural network model to classify second synthetic images generated using the synthetic image generation model as either real-like or non-real-like; and
a target model training component that employs a subset of the second synthetic images classified as real-like by the inference component to train a machine learning model to perform artificial intelligence analytics on real images.
2. The system of claim 1 , wherein the real images, the first real images, the synthetic images, the first synthetic images and the second synthetic images comprise medical images.
3. The system of claim 2 , wherein the artificial intelligence analytics comprise automated evaluation of one or more medical conditions reflected in the real images.
4. The system of claim 1 , wherein the synthetic image generation model comprises a generative adversarial network model.
5. The system of claim 1 , wherein the computer executable components further comprise:
a synthetic image generator that generates the synthetic images using the synthetic image generation model, the synthetic images comprising the first synthetic images and the second synthetic images.
6. The system of claim 1 , wherein the deep neural network model comprises a very deep convolutional neural network model configured for binary classification.
7. A method, comprising:
training, by a system operatively coupled to a processor, a deep neural network model to classify synthetic images generated using a synthetic image generation model as either real-like or non-real like using a training dataset comprising first real images and first synthetic images generated using the synthetic image generation model and classified as non-real like, resulting in a trained deep neural network model;
classifying, by the system, second synthetic images generated using the synthetic image generation model as either real-like or non-real-like using the trained deep neural network model; and
employing, by the system, a subset of the second synthetic images classified as real-like to train a machine learning model to perform artificial intelligence analytics on real images.
8. The method of claim 7 , wherein the real images, the first real images, the synthetic images, the first synthetic images and the second synthetic images comprise medical images.
9. The method of claim 8 , wherein the artificial intelligence analytics comprise automated evaluation of one or more medical conditions reflected in the real images.
10. The method of claim 7 , wherein the synthetic image generation model comprises a generative adversarial network model.
11. The method of claim 7 , further comprising:
generating, by the system, the synthetic images using the synthetic image generation model, the synthetic images comprising the first synthetic images and the second synthetic images.
12. The method of claim 7 , wherein the deep neural network model comprises a very deep convolutional neural network model configured for binary classification.
13. A machine-readable storage medium, comprising executable instructions that, when executed by a processor, facilitate performance of operations, comprising:
training, by a system operatively coupled to a processor, a deep neural network model to classify synthetic images generated using a synthetic image generation model as either real-like or non-real like using a training dataset comprising first real images and first synthetic images generated using the synthetic image generation model and classified as non-real like, resulting in a trained deep neural network model;
classifying second synthetic images generated using the synthetic image generation model as either real-like or non-real-like using the trained deep neural network model; and
employing a subset of the second synthetic images classified as real-like to train a machine learning model to perform artificial intelligence analytics on real images.
14. The machine-readable storage medium of claim 13 , wherein the real images, the first real images, the synthetic images, the first synthetic images and the second synthetic images comprise medical images.
15. The machine-readable storage medium of claim 14 , wherein the artificial intelligence analytics comprise automated evaluation of one or more medical conditions reflected in the real images.
16. The machine-readable storage medium of claim 13 , wherein the synthetic image generation model comprises a generative adversarial network model.
17. The machine-readable storage medium of claim 13 , wherein the operations further comprise:
generating synthetic images using the synthetic image generation model, the synthetic images comprising the first synthetic images and the second synthetic images.
18. The machine-readable storage medium of claim 13 , wherein the deep neural network model comprises a very deep convolutional neural network model configured for binary classification.