Skeleton-Guided Image Generation for Precise Human Pose Control

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image generation AI systems struggle to finely designate the posture of a person in generated images, limiting the flexibility and precision of image creation.

Innovation Solution

An image generation device that extracts a skeleton feature value from skeleton information and uses a machine learning model to generate images based on this feature value, removing noise stepwise to produce clear images with designated postures.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If text input is used for image generation, then the system is simple to operate, but the posture of the person cannot be finely designated

Engineering Contradiction:

Improveease of operationVSAvoidposture designation precision

Core Design Contradiction:

Ease of operationVSManufacturing precision

Solution Approach 1:

The patent introduces skeleton information as an intermediary between the user's posture designation intent and the image generation process. The skeleton information serves as a mediator that translates desired posture into a format that can be processed by the machine learning model, enabling precise posture control while maintaining ease of operation through intuitive skeleton-based input.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent changes the input parameter from text to skeleton information, which includes positional data of joints and bones. This parameter change enables the system to directly represent and control posture information, transforming the image generation process into one that can precisely designate postures while remaining user-friendly through visual skeleton-based interaction.

Inventive Principle:

Principle #35Parameter changes

2Manufacturing precision

If skeleton information is used for image generation, then posture designation precision is improved, but the system complexity increases

Engineering Contradiction:

Improveposture designation precisionVSAvoidsystem complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent uses a machine learning model that has been trained to recognize and process skeleton information patterns. Instead of implementing complex custom processing logic, the system leverages a pre-trained model that copies and adapts existing knowledge about human skeletons, thereby achieving precise posture designation without proportionally increasing system complexity.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The machine learning model performs self-learning during the training phase to automatically capture the relationships between skeleton information and corresponding images. This self-service capability allows the system to handle complex posture variations autonomously without requiring manual programming of every possible scenario, thus managing system complexity while maintaining high precision.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20260044996A1Image generation device, image generation method, and computer-readable recording medium

Publication Date: 2026.02.12 NEC CORP

AI summary

An image generation device includes a feature value extraction unit that extracts a skeleton feature value of a skeleton from skeleton information specifying a position of each of joints constituting the skeleton, and an image generation unit that generates an image according to the skeleton by inputting the extracted skeleton feature value and a noise image to a machine learning model for estimating noise that has been added to the image and removing the noise from the image using an output result from the machine learning model.