Image Feature Label Editing for Consistent AI Scene Updates

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video game development processes are time-consuming and inefficient when making changes to image characteristics, as they often require intricate steps and can lead to inconsistencies across objects within a scene, necessitating redraws or wholesale changes.

Innovation Solution

An image generation AI model implementing latent diffusion techniques is used to modify labeled features of an image, allowing users to edit and classify features through an intuitive interface, with user feedback guiding the generation of a modified image that maintains consistency and reduces the need for redraws.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If traditional manual methods are used to modify image characteristics, then developers can make precise changes to individual parameters, but the process becomes extremely time-consuming and requires intricate steps

Engineering Contradiction:

Improveprecision of parameter modificationVSAvoidtime to modify image characteristics

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent replaces manual mechanical editing processes with an AI-based automated system. The AI model automatically identifies objects, extracts features, and generates modified images based on natural language instructions, eliminating the need for developers to manually adjust each parameter through intricate steps, thus resolving the contradiction between precision and time consumption

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces an AI system as an intermediary between the developer's intent and the final image modification. This intermediary automatically interprets natural language instructions, identifies relevant image features, and applies precise modifications without requiring the developer to manually navigate complex editing parameters, thereby reducing time while maintaining precision

Inventive Principle:

Principle #24Intermediary (Mediator)

2Ease of manufacture

If developers manually adjust parameters of object characteristics, then specific changes can be made to individual objects, but consistency with other objects in the scene is difficult to maintain

Engineering Contradiction:

Improveease of making specific changesVSAvoidconsistency across objects in scene

Core Design Contradiction:

Ease of manufactureVSStability of the object's composition

Solution Approach 1:

The patent implements a universal AI-based system that can consistently apply modifications across multiple objects and the entire scene. The system analyzes global scene characteristics and ensures that changes to individual objects maintain consistency with other objects, providing both ease of making specific changes and stability of scene composition simultaneously

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Adaptability or versatility

If wholesale changes are made to the scene environment, then the entire scene can be updated, but it requires redrawing the image from scratch

Engineering Contradiction:

Improveability to make wholesale changesVSAvoidtime to redraw entire image

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent segments the image modification process into targeted regions and objects rather than requiring complete redrawing. The AI system identifies specific areas needing modification and applies changes only to those segments while preserving the rest of the scene, enabling wholesale environmental changes without the need to redraw the entire image from scratch

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary analysis of the scene to identify objects, features, and relationships before making modifications. This preliminary action allows the system to plan and execute wholesale changes efficiently by understanding the scene structure in advance, avoiding the need for complete redrawing and significantly reducing the time required for major scene updates

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12530913B2Qualifying labels automatically attributed to content in images

Publication Date: 2026.01.20 SONY INTERACTIVE ENTERTAINMENT LLC

AI summary

A method for image generation. The method including identifying a plurality of features of an image. The method including classifying each of the plurality of features using an artificial intelligence (AI) model trained to identify features in a plurality of images, wherein the plurality of features is classified as a plurality of labels, wherein the image is provided as input to the AI model. The method including receiving feedback for a label, wherein the feedback is associated with a user. The method including modifying a label based on the feedback. The method including updating the plurality of labels with the label that is modified. The method including providing as input the plurality of labels that is updated into an image generation artificial intelligence system configured for implementing latent diffusion to generate an updated image.