/Patent Tracker
SonyPatent applicationAI & softwareInterest 6/10

Sony patents tool to spot flaws in AI-generated video frames

Patent drawing
Patent drawing from the official filing.

Sony has filed a patent for a system that checks for visual errors in video frames created by machine learning models. The tool compares a newly generated frame against a reference image to identify pixel-level differences, helping developers find defects in the output.

Why it mattersThis is interesting for tech workers because it offers a specific method for quality control in AI video generation, ensuring that synthetic images are accurate and free of visual glitches.

In the filing’s own words

To make it possible to easily check for an image defect in an estimated frame ( 24) output from a machine learning model ( 200). An image defect inspector ( 100) including at least one processor, wherein the at least one processor: acquires an n-th (where n is a natural number of 2 or greater) estimated frame ( 24) output from a machine learning model ( 200) by an n-th input frame ( 22) and (n−1)-th auxiliary information ( 28) related to features of first to (n− 1)-th input frames ( 22) being input to the machine learning model ( 200); acquires a predetermined comparison target frame; detects a difference related to pixel information of the n-th estimated frame ( 24) with respect to the comparison target frame, based on the n-th estimated frame ( 24) and the comparison target frame; and acquires image defect information of the n-th estimated frame ( 24), based on the difference related to the pixel information.

What exactly they claim (claim 1)

1. An image inspection system comprising: one or more processors; and one or more memories storing executable instructions thereon that, upon execution by the one or more processors, cause the image inspection system to at least acquire an n-th estimated frame output from a machine learning model, the n-th estimated frame being generated by the machine learning model by inputting an n-th input frame and (n−1)-th auxiliary information related to features of first to (n−1)-th input frames into the machine learning model; acquire a predetermined comparison target frame; detect a difference related to pixel information of the n-th estimated frame with respect to the predetermined comparison target frame, based on the n-th estimated frame and the predetermined comparison target frame; and acquire image defect information of the n-th estimated frame, based on the difference related to the pixel information, wherein n is a natural number greater than 1.

Facts

Company
Sony (Sony Interactive Entertainment Inc.)
Official title
IMAGE INSPECTION SYSTEM, IMAGE INSPECTION METHOD, AND PROGRAM
Publication number
US20260311314A1
Application number
19638957
Filed
Apr 3, 2026
Published
Oct 8, 2026
Inventors
1
Claims
10

Classified under

G06T 7/0002 Physics & computingG06T 7/11 Physics & computingG06T 7/97 Physics & computingG06T 2207/10016 Physics & computingG06T 2207/20081 Physics & computing

More from Sony in AI & software

Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.