USPTO Patent Grant: Synthetic-to-realistic Image Conversion using GANs

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 24th, 2026

Detected March 25th, 2026

Summary

The USPTO has granted patent US12586359B2 to Raytheon Company for a method of synthetic-to-realistic image conversion using generative adversarial networks (GANs) or other machine learning models. The patent covers training methods for improving image generation quality.

View original document View source feed page

What changed

The United States Patent and Trademark Office (USPTO) has issued patent US12586359B2 to Raytheon Company, detailing a method for synthetic-to-realistic image conversion utilizing generative adversarial networks (GANs) or other machine learning models. The patent describes a process of training machine learning models with image pairs, including initial training data with high-quality images and subsequent training with lower-quality or less aligned images to enhance the model's ability to generate realistic outputs.

This patent grant is primarily an intellectual property matter and does not impose direct regulatory obligations or compliance deadlines on regulated entities. However, it signifies advancements in AI and machine learning technology, particularly in image generation, which may be relevant for companies operating in sectors that utilize or develop such technologies. Compliance officers in technology and manufacturing sectors should be aware of patent grants in AI as they can impact intellectual property strategies and competitive landscapes.

Source document (simplified)

← USPTO Patent Grants

Synthetic-to-realistic image conversion using generative adversarial network (GAN) or other machine learning model

Grant US12586359B2 Kind: B2 Mar 24, 2026

Assignee

Raytheon Company

Inventors

Jonathan H. Goldstein, Lauren E. Turney, Richard W. Ely, Jody D. Verret, Brett N. Appleton

Abstract

A method includes obtaining training data having first image pairs, where each of the first image pairs includes (i) a first training image and (ii) a first ground truth image. The method also includes training a machine learning model to generate realistic images using the first image pairs. The method further includes obtaining additional training data having second image pairs, where each of the second image pairs includes (i) a second training image and (ii) a second ground truth image. At least some of the images in the second image pairs are less aligned or of lower quality than at least some of the images in the first image pairs. In addition, the method includes continuing to train the machine learning model to generate the realistic images using the second image pairs.