Local Attribute Image Editing Using Mask-Guided Model Fusion

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image editing technologies, such as PhotoShop, require manual editing by professionals, leading to inefficient image editing processes.

Innovation Solution

Utilizing an initial image generation model trained on a first training image set and a feature image generation model trained on a second training image set, combined with attribute mask images, to automatically edit specific attributes in images without altering other features.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If manual editing using image editing software is used, then editing precision is improved, but productivity deteriorates

Engineering Contradiction:

Improveediting precisionVSAvoidproductivity

Core Design Contradiction:

Manufacturing precisionVSProductivity

Solution Approach 1:

The patent replaces the mechanical manual editing process with an automated image processing system that uses computer vision algorithms and machine learning models to perform attribute editing tasks, thereby substituting human manual operations with automated computational processes

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service image editing by automatically analyzing image content, identifying target attributes, and performing editing operations without requiring professional manual intervention, allowing the system to serve itself in completing the editing workflow

Inventive Principle:

Principle #25Self-service

2Manufacturing precision

If professional manual editing is used, then editing quality is improved, but loss of time increases

Engineering Contradiction:

Improveediting qualityVSAvoidloss of time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The system performs preliminary actions by pre-training machine learning models on large datasets and pre-processing images through automatic content analysis and attribute identification, so that when editing is needed, the heavy computational work has already been prepared in advance

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces time-consuming manual editing operations with automated computer vision systems that can process and edit images computationally in seconds, substituting the mechanical process of manual pixel-level editing with algorithmic image processing

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS12548210B2Local attribute image editing using an image generation model and a feature image generation model

Publication Date: 2026.02.10 TENCENT TECHNOLOGY (SHENZHEN) CO LTD

AI summary

An image editing method includes acquiring an initial image generation model and a feature image generation model, the initial image generation model having been trained based on a first training image set, the feature image generation model having been obtained by training the initial image generation model based on a second training image set. The method further includes acquiring a joint mask image based on image regions corresponding to the target attribute in the object images, and acquiring a second initial object image and a second feature object image output by corresponding target network layers. The method further includes fusing the second initial object image and the second feature object image based on the joint mask image to obtain a reference object image.