Samsung Granted Patent for Image Orientation Correction

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 24th, 2026

Detected March 25th, 2026

Summary

The USPTO has granted Samsung Electronics Co., Ltd. a patent for a method and system for image orientation correction using machine learning trained via federated learning. This technology aims to improve image tilt correction while maintaining user data privacy.

View original document View source feed page

What changed

The United States Patent and Trademark Office (USPTO) has granted Samsung Electronics Co., Ltd. patent US12586156B2 for an "Method, system and apparatus for image orientation correction." The patent covers a computer-implemented method for generating a training dataset for a machine learning (ML) model using federated learning to perform image orientation correction. This approach aims to create a diverse training dataset that accounts for various image tilt angles while preserving user data privacy.

This patent grant is primarily of interest to technology companies and manufacturers involved in image processing and AI development. While it does not impose new regulatory obligations on businesses, it represents a technological advancement in the field of AI-driven image correction. Compliance officers should note this development as it pertains to intellectual property in AI and data privacy techniques, particularly federated learning, which is a key aspect of the patented technology.

Source document (simplified)

← USPTO Patent Grants

Method, system and apparatus for image orientation correction

Grant US12586156B2 Kind: B2 Mar 24, 2026

Assignee

SAMSUNG ELECTRONICS CO., LTD.

Inventors

Albert Saa-Garriga, Mehmet Yucel, Tommaso Maestri, Konstantinos Peppas

Abstract

Broadly speaking, the present techniques generally relates to methods, systems and apparatuses for performing image orientation correction, i.e. correcting or adjusting images that are tilted. In particular, the present application relates to a computer-implemented method for generating a training dataset for training a machine learning, ML, model using federated learning to perform image orientation correction, and methods for training the ML model using the generated training dataset. Advantageously, the method to generate a training dataset enables a diverse training dataset to be generated while maintaining user data privacy, where the diversity refers to the range of image tilt angles represented by the dataset. The present application also provides methods for training the ML model using the generated training dataset.

CPC Classifications

G06N 3/044 G06N 3/045 G06N 3/0464 G06N 3/047 G06N 3/0475 G06N 3/092 G06N 3/094 G06N 3/098 G06T 3/60 G06V 10/774 G06V 10/776 G06V 10/82 G06V 10/95 G06V 2201/10