US 10,540,798 B1Granted

Methods and arrangements to create images

AssigneeCapital One Services, LLC

InventorsAustin Grant Walters, Fardin Abdi Taghi Abad, Jeremy Edward Goodsitt, Reza Farivar, Kenneth Taylor

Life of this patent

Estimating the term…

Loading patent…

Abstract

Logic may create new images by adding more than one synthetic image on a template. Logic may provide a template with a background for a new image. Logic may provide a set of models, each model to comprise a generative adversarial network (GAN), the GANs pretrained independently to generate different synthetic images. Logic may select two or more models from the set of models. Logic may generate, by the two or more models, two or more of the different synthetic images. Logic may combine the two or more of the different synthetic images with the template to create the new image. And logic may train a set of GANs independently, to generate one of two or more different synthetic images on a blank image background, the different synthetic images to comprise a subset of the multiple features of a new image.

Claims20

Dependent claims

An apparatus comprising:

memory;
and
logic circuitry coupled with the memory to
provide a template, the template to comprise a background for a new image;
provide a set of models, each model to comprise a generative adversarial network (GAN), the GANs pretrained independently to generate different synthetic images;
select two or more models from the set of models;
generate, by the two or more models, two or more of the different synthetic images;
determine an order to combine the two or more of the different synthetic images with the template, the determined order established randomly or by using one or more parameters associated with the different synthetic images;
and
combine the two or more of the different synthetic images with the template in the determined order to create the new image, wherein the combination of each of the two or more of the different synthetic images with the template comprises:
use of a first model to generate a first synthetic image of the different synthetic images on the template and output the first synthetic image;
and
use of the remaining of the two or more models in accordance with the determined order, wherein the second model of the set receives the first synthetic image as an input, wherein each of the remaining of the two or more models generates an output of a subsequent synthetic image of the different synthetic images on the input synthetic image, wherein the output of each of the remaining of the two or more models is provided as an input to a successive model of the remaining of the two or more models in accordance with the determined order, and wherein an image output by a last model of the two or more models is the new image.

10 more elements

A non-transitory storage medium containing instructions, which when executed by a processor, cause the processor to perform operations, the operations to:

provide a base image, the base image to comprise a background for a new image;
provide a library of models, each model to comprise a generative adversarial network (GAN), the GAN pretrained to generate a different synthetic image;
select two or more models from the library of models;
generate, by the two or more models, two or more of the different synthetic images;
and
determine an order by an overlay model to combine the two or more of the different synthetic images with a template, the determined order established randomly or by using one or more parameters associated with the different synthetic images;
and
combine the two or more of the different synthetic images with the base image in the determined order to create the new image, wherein the combination of each of the two or more of the different synthetic images with the template comprises:
use of a first model to generate a first synthetic image of the different synthetic images on the template and output the first synthetic image;
and
use of the remaining of the two or more models in accordance with the determined order, wherein the second model of the set receives the first synthetic image as an input, wherein each of the remaining of the two or more models generates an output of a subsequent synthetic image of the different synthetic images on the input synthetic image, wherein the output of each of the remaining of the two or more models is provided as an input to a successive model of the remaining of the two or more models in accordance with the determined order, and wherein an image output by a last model of the two or more models is the new image.

8 more elements

Dependencies are inferred from the claim text.