2D Virtual Clothing Fitting via Hybrid Deep Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional systems for two-dimensional (2D) virtual clothing fitting face challenges in realism and accuracy due to manual labor costs and limitations in scaling, as well as lower accuracy and authenticity when using basic machine learning models.

Innovation Solution

A hybrid deep learning technology integrating optimization and deterministic classification algorithms is used to reconstruct user images combined with clothing images, transforming clothing images to fit the user's size through machine learning models and optimization algorithms, comprising four main blocks: data preprocessing, shape modification, swapping, and calibration and optimization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual work of simulation graphics engineers is used to reconstruct user images on clothing images, then the realism and accuracy of the result is improved, but the labor cost and implementation time are extremely huge and cannot be expanded on an industrial scale

Engineering Contradiction:

Improveaccuracy of user size descriptionVSAvoidprocessing speed and scalability

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces manual mechanical editing operations with automated machine learning models. Specifically, it uses a deep learning-based system that automatically segments users and clothing, reconstructs user images, and transforms clothing images to fit user dimensions, eliminating the need for manual graphics engineering while maintaining high accuracy and enabling industrial-scale deployment

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes the operational parameters from manual pixel-level editing to automated algorithmic processing. It uses optimization algorithms to adjust clothing image parameters (scale, position, shape) to match user dimensions, transforming the process from labor-intensive manual adjustment to automated parameter optimization that scales efficiently

Inventive Principle:

Principle #35Parameter changes

2Productivity

If basic machine learning models are used to directly graft user face onto model image, then the processing speed is improved, but the realism and authenticity of the result deteriorates

Engineering Contradiction:

Improveprocessing speedVSAvoidrealism and authenticity of user representation

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments the complex task into multiple specialized modules: user segmentation, clothing segmentation, user image reconstruction, and clothing image transformation. Each module handles a specific aspect independently, allowing basic ML models to process each segment efficiently while the integrated system maintains high realism through coordinated processing of all segments

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system introduces intermediate processing steps including optimization algorithms that mediate between the user image and clothing image. These intermediaries (optimization algorithms, transformation models) refine the direct grafting process to maintain realism while preserving the speed benefits of automated processing

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20250111426A1System for two-dimensional (2D) virtual clothing fitting using a hybrid deep learning technology integrating optimization and deterministic classification algorithms

Publication Date: 2025.04.03 VIETTEL GRP

AI summary

The invention relates to a two-dimensional (2D) virtual clothing fitting system using machine learning technology and a deterministic classification algorithm. The system allows the construction of a two-dimensional (2D) digital interactive image between a cloth and the user's body instead of using a graphic engineer to simulate it. Machine learning models of image processing combined with a deterministic classification algorithm reproduce the user's image combined with the garment image and transform the cloth image according to the user's size.