Pairwise Avatar Image Generation Using Pre-Trained Models

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing painting software programs can only generate single images and are unable to create images with pairwise relationships, such as couple or bestie images, failing to meet the need for personalized avatars with specific relationships.

Innovation Solution

An image generation method that involves obtaining target information including a text set, transmitting an image generation request to a server, and receiving M groups of images generated by the server based on the target information and a pre-trained image generation model, where each group includes two images with a pairwise relationship.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If existing painting software programs are used to generate personalized avatars, then single image generation is achieved, but images with pairwise relationships (such as couple images and bestie images) cannot be generated

Engineering Contradiction:
Improveimage relationship typeVSAvoidsoftware functionality
Core Design Contradiction:
Adaptability or versatilityVSEase of manufacture

Solution Approach 1:

The image generation model is designed to handle multiple types of image generation tasks including single image generation and pairwise relationship image generation (couple images, bestie images). The model takes text descriptions and generates appropriate images based on the relationship type, making the system versatile and adaptable to different user needs without requiring separate software for each function.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system segments the image generation process into distinct stages: receiving text description, identifying relationship type, generating appropriate images based on relationship category, and presenting results. This segmentation allows the system to handle different relationship types systematically while maintaining overall functionality.

Inventive Principle:
Principle #1Segmentation

2Adaptability or versatility

If a pre-trained image generation model is used to generate M groups of images with pairwise relationships, then personalized avatar requirements are met, but the complexity of the generation system increases

Engineering Contradiction:
Improvepersonalized avatar generationVSAvoidsystem structure
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The image generation model is pre-trained with knowledge of various pairwise relationships and image generation patterns before being deployed. This preliminary training allows the system to handle complex relationship-based image generation tasks without requiring complex real-time processing logic, thereby reducing operational system complexity while maintaining high adaptability for personalized avatar generation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses a pre-trained model that has learned from vast datasets of images and their relationships. By leveraging this pre-existing knowledge and patterns stored in the model, the system can generate personalized avatar images without requiring complex computation during actual use, simplifying the operational system structure.

Inventive Principle:
Principle #26Copying

Data Source

PatentUS20250292451A1Image generation method, apparatus, and device, and storage medium
Publication Date: 2025.09.18 BEIJING SOGOU TECHNOLOGY DEVELOPMENT CO LTD
  • US20250292451A1 patent drawing
  • US20250292451A1 patent drawing
  • US20250292451A1 patent drawing

AI summary

This application provides an image generation method, apparatus, and device, and a storage medium, and relates to the field of computer technologies. The method includes: obtaining target information for performing image generation, the target information including a text set; transmitting an image generation request to a server in response to an image generation operation on the target information, the image generation request carrying the target information; receiving M groups of images transmitted by the server, each group of the M groups of images including two images that have a pairwise relationship in terms of preset content, the M groups of images being generated by the server based on the target information and a pre-trained image generation model, and M being a positive integer; and displaying the M groups of images.