Object Relationship Graph Editing for AI Image Modification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image editing technologies lack the ability to automatically apply image effects based on the relationships between objects in an image, leading to suboptimal editing results.
Innovation Solution
An artificial intelligence system that utilizes a Graph Neural Network (GNN) model to analyze object relationships and apply image effects accordingly, generating a modified image that enhances the overall image quality by considering the interactions and features of the objects within the image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If rule-based systems are used for image editing, then editing parameters can be manually controlled, but the system cannot automatically learn and adapt to user preferences
Solution Approach 1:
The patent replaces rule-based mechanical systems with an AI-based learning system. The system uses machine learning models to automatically learn user preferences from interaction data, substituting manual rule configuration with automated intelligent decision-making that adapts to individual user needs.
Solution Approach 2:
The AI system performs self-learning and self-improvement by automatically analyzing user interactions and updating its models without requiring manual reconfiguration. The system serves itself by continuously improving its understanding of user preferences through accumulated data.
2Extent of automation
If AI systems are used for image editing, then automatic learning and judgment are achieved, but the system lacks understanding of relationships between multiple objects in the image
Solution Approach 1:
The patent segments the image into multiple objects and establishes relationship graphs connecting them. Each object is analyzed individually, and their relationships are modeled separately, allowing the AI to understand both individual object characteristics and their contextual relationships within the image.
Solution Approach 2:
The patent adds a relational dimension to the analysis by creating relationship graphs that connect objects. This transforms the problem from analyzing individual objects in isolation to analyzing a multi-dimensional structure where objects and their relationships are jointly considered, enabling more sophisticated editing decisions.
3Ease of manufacture
If uniform image effects are applied to the entire image, then processing is simple, but the editing results do not suit different objects within the image
Solution Approach 1:
The patent applies different image effects to different objects based on their specific characteristics and relationships. Instead of uniform processing, the system determines appropriate effects locally for each object, such as applying different filters, adjustments, or enhancements tailored to the specific needs of each detected object within the image.
Data Source
AI summary
An electronic device includes: a processor; and a memory storing instructions. By executing the instructions, the processor is configured to: receive a first image, recognize a plurality of objects in the first image to generate object information representing the plurality of objects, generate an object relationship graph including relationships between the plurality of objects, based on the first image and the object information, obtain image effect data including image effects to be respectively applied to the plurality of objects by inputting the object relationship graph to an image modification Graph Neural Network (GNN) model, and generate a modified image based on the first image, the object information, and the image effect data.