Generative Model Image Expansion for Object Movement

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing technologies struggle to maintain user satisfaction when objects are transformed within an image, as partial objects or mismatched objects can result from size or position changes.

Innovation Solution

A method using generative models to process images by receiving user input for object movement, expanding the image in the determined direction, determining a generation required area, generating an image for this area using a generative model, and outputting a recomposed image.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If the position or size of an object is transformed in the image, then the user can edit the image according to their needs, but partial objects or mismatched objects are displayed resulting in decreased user satisfaction

Engineering Contradiction:
Improveimage editing capabilityVSAvoidobject transformation quality
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system performs preliminary actions by expanding the image in the direction opposite to the object movement before generating the final recomposed image. This preliminary expansion creates the necessary space and context for the object transformation, ensuring that the object remains properly framed and matches the surrounding environment after the transformation is applied.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary process using generative models to create a partial image that bridges the original image and the transformed object. This intermediary partial image is generated based on the expanded image and the transformation parameters, serving as a mediator that ensures seamless integration of the transformed object while maintaining consistency with the surrounding environment.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If the image is expanded to accommodate object movement, then the object can be positioned or sized as required, but the processing time and computational resources increase

Engineering Contradiction:
Improveobject transformation flexibilityVSAvoidprocessing time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system applies local quality by expanding the image only in the specific direction opposite to the object movement rather than expanding the entire image uniformly. This localized expansion approach minimizes the computational resources and processing time required while still providing the necessary space for the object transformation.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent employs partial action by generating only the necessary partial image for the generation required area rather than processing the entire image. The partial image is generated based on the expanded image and the specific transformation parameters, reducing the computational burden while maintaining the necessary adaptability for object transformation.

Inventive Principle:
Principle #16Partial or excessive action

3Manufacturing precision

If generative models are used to generate the partial image, then the image quality and style consistency are improved, but the computational complexity and model requirements increase

Engineering Contradiction:
Improveimage quality consistencyVSAvoidmodel requirements
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The system segments the image processing task into distinct components: expanding the image in the opposite direction, determining the generation required area, and generating only the necessary partial image using the generative model. This segmentation allows the complex generative model to be applied only to the relevant portion of the image, reducing the overall computational complexity while maintaining image quality consistency.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS20250045877A1Image processing method using generative model and computing device for performing the same
Publication Date: 2025.02.06 SAMSUNG ELECTRONICS CO LTD
  • US20250045877A1 patent drawing
  • US20250045877A1 patent drawing
  • US20250045877A1 patent drawing

AI summary

A method of editing an image by using a generative model includes receiving a user input for a movement of at least one object included in an input image; expanding the input image in a direction based on the movement of the at least one object; determining a generation required area based on the expanded input image, wherein the generation required area is an area in which generation of a partial image for the at least one object is required; generating the partial image for the at least one object in the generation required area by using at least one generative model; and outputting a recomposed image based on the input image and the partial image for the at least one object.