System and method for lossy image and video compression utilizing a metanetwork

• US 10,489,936 B1
• Filed: 05/16/2019
• Issued: 11/26/2019
• Est. Priority Date: 04/29/2019
• Status: Active Grant

• Current Assignee: Deep Render Ltd.
• Original Assignee: Deep Render Ltd.
• Inventors: Zafar, Arsalan Ali, Besenbruch, Christian Lars
• Primary Examiner(s): Nakhjavan, Shervin K
• Application Number:
  US16/413,770
• Time in Patent Office: 194 Days
• Field of Search: None
• CPC Class Codes: G06N 20/20 Ensemble learning
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/063 using electronic means
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06T 3/4046 using neural networks
• CPC Class Codes: G06T 3/4084 in the transform domain, e.g. fast Fourier transform [FFT] domain scaling
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06T 9/002 using neural networks
• CPC Class Codes: H04N 19/90 using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals

Abstract

A system and method for lossy image and video compression that utilizes a metanetwork to generate a set of hyperparameters necessary for an image encoding network to reconstruct the desired image from a given noise image.

Claims

1. A system for lossy image and video compression utilizing a metanetwork, comprising:

a metanetwork engine comprising a processor, a memory, and a first plurality of programming instructions stored in the memory, wherein the first plurality of programming instructions, when operating on the processor, cause the processor to;

receive a desired image;

receive a noise image;

receive a set of training images;

using the set of training images, train a plurality of neural networks to reconstruct each of the set of training images by mapping the noise image to each of the set of training images;

store the parameters for each of the plurality of neural networks as a set of metanetwork hyperparameters;

use the set of metanetwork hyperparameters as operating parameters for each of the plurality of neural networks;

use the plurality of neural networks to map the noise image to the desired image, producing a second set of hyperparameters corresponding to the specific filters produced from the operation of each of the plurality of neural networks, such that the second set of hyperparameters, when applied to the noise image using the neural network, produce an approximation of the desired image within an error that is less than a pre-determined threshold; and

store the second set of hyperparameters for use in future image mapping operations.

2. The system of claim 1, wherein each of the plurality of neural networks:

generates at least one convolutional filter, wherein the noise image may be filtered through all convolutional filters in succession, mapping it to an approximation of a desired image; and

facilitates communication between the plurality of neural networks to alleviate the vanishing gradient problem.

3. The system of claim 2, wherein the plurality of neural networks may be located on separate computing devices, connected across a network.

4. The system of claim 1, wherein the noise image is static and unchanging.

5. A method for lossy image compression utilizing a metanetwork, comprising the steps of:

receiving a desired image;

receiving a noise image;

receiving a set of training images;

using the set of training images to train a plurality of neural networks to reconstruct each of the set of training images by mapping the noise image to each of the set of training images;

storing the parameters for each of the plurality of neural networks as a set of metanetwork hyperparameters;

using the set of metanetwork hyperparameters as operating parameters for each of the plurality of neural networks;

using the plurality of neural networks to map the noise image to the desired image, producing a second set of hyperparameters corresponding to the specific filters produced from the operation of each of the plurality of neural networks, such that the second set of hyperparameters, when applied to the noise image using the neural network, produce an approximation of the desired image within an error that is less than a pre-determined threshold; and

storing the second set of hyperparameters for use in future image mapping operations.

6. The method of claim 5, further comprising the steps of:

Generating, at each of the plurality of neural networks, at least one convolutional filter, wherein a noise image may be filtered through the convolutional filters in succession, mapping it to an approximation of a desired image, using a plurality of neural meta-networks; and

facilitating communication between the plurality of neural networks to alleviate the vanishing gradient problem.

7. The method of claim 6, wherein the plurality of neural networks may be located on separate computing devices, connected across a network.

8. The method of claim 5, wherein the noise image is static and unchanging.