AI Device and 3D Agency Generating Method

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 31st, 2026

Detected April 1st, 2026

Summary

USPTO granted patent US12592060B2 to LG Electronics Inc. on March 31, 2026, covering an artificial intelligence device and method for generating three-dimensional agency from two-dimensional images. The patent includes 13 claims and encompasses CPC classifications related to neural networks, computer graphics, and voice recognition technology.

View original document View source feed page

What changed

USPTO issued Patent US12592060B2 to LG Electronics Inc. for an AI device and 3D agency generating method that acquires 2D images of a target persona, extracts training data including voice, face, and motion information, trains an agency generation model, and generates a 3D agency expressing the target persona's motion features upon user request. The patent names Taeju Hwang, Jongjin Park, Yubin Yoon, Taehwan Hwang, and Hyoyoung Kim as inventors, with Application No. 18199017 filed May 18, 2023.

This is a routine patent grant providing intellectual property protection to LG Electronics Inc. No immediate compliance action is required by third parties. Competitors developing similar AI-based 3D avatar or digital persona technology should review the patent claims to assess potential infringement risks and freedom-to-operate implications. The 13 claims define the scope of protected subject matter under CPC classifications G06N 3/08, G06T 13/40, and various G06V and G10L categories.

Source document (simplified)

← USPTO Patent Grants

Artificial intelligence device and 3D agency generating method thereof

Grant US12592060B2 Kind: B2 Mar 31, 2026

Assignee

LG ELECTRONICS INC.

Inventors

Taeju Hwang, Jongjin Park, Yubin Yoon, Taehwan Hwang, Hyoyoung Kim

Abstract

A three-dimensional agency generating method of an artificial intelligence device according to an embodiment of the present disclosure may include: acquiring a two-dimensional image including a target persona; pre-processing the two-dimensional image to extract a training dataset including voice information, face information, and motion information corresponding to the target persona; inputting the training dataset to an agency generation model to learn motion features of the target persona; confirming whether user input data for generating the three-dimensional agency is received; and when the user input data for generating the three-dimensional agency is received, inputting the user input data to the pre-trained agency generation model to generate the three-dimensional agency expressing the motion features of the target persona.