Virtual Fabric Drape Parameter Estimation With Neural Network Feedback

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for estimating physical property parameters in virtual garment simulation lack accuracy and efficiency, particularly in reproducing the drape of fabrics, necessitating improved methods to enhance precision and speed.

Innovation Solution

A method utilizing a neural network trained on the correlation between physical property parameters and a mesh of draped virtual fabric, combined with an optimization process to update parameters based on error analysis, to generate accurate drape simulations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional mechanical measurement methods are used to obtain physical property parameters, then measurement precision is improved, but productivity deteriorates due to time-consuming manual processes

Engineering Contradiction:

Improveaccuracy of physical property parametersVSAvoidspeed of parameter acquisition

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces traditional mechanical measurement devices with a neural network-based learning model that processes image data to estimate physical property parameters. The system uses computer vision and machine learning algorithms to automatically extract fabric properties from images, eliminating manual mechanical measurement while maintaining or improving accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent creates a virtual copy of the fabric measurement process through a neural network model trained on paired image-parameter data. Instead of physically measuring each fabric sample, the system learns from training data to predict parameters from images, enabling rapid estimation without repeated mechanical measurements.

Inventive Principle:

Principle #26Copying

2Manufacturing precision

If optimization methods are used to minimize error between simulation and target results, then manufacturing precision is improved, but productivity deteriorates due to computational complexity

Engineering Contradiction:

Improveaccuracy of drape simulationVSAvoidcomputation time for parameter optimization

Core Design Contradiction:

Manufacturing precisionVSProductivity

Solution Approach 1:

The patent performs preliminary training of the neural network model using optimization algorithms during the offline training phase. The model learns optimal parameter mappings from training data, so that during actual use, parameter estimation occurs rapidly through forward propagation without requiring repeated optimization computations for each new fabric.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent transforms the optimization problem from an online computational task to an offline training task. By changing the parameters (neural network weights) during training through optimization, the system creates a pre-optimized model that can quickly estimate parameters for new fabrics without requiring real-time optimization computations.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If learning models are used to estimate physical property parameters, then productivity is improved through rapid computation, but measurement precision deteriorates due to approximation errors

Engineering Contradiction:

Improvespeed of parameter estimationVSAvoidaccuracy of parameter estimation

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent implements a feedback mechanism where the neural network's parameter estimates are used in drape simulation, and the simulation results are compared with target drape data. The estimated parameters are then updated based on the error between simulated and target results, creating a closed-loop system that continuously improves estimation accuracy.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent combines multiple components into a composite estimation system: a neural network for initial parameter estimation, an error calculation module for comparing simulation with target, and an parameter update mechanism. This composite approach leverages the speed of neural networks while correcting accuracy through iterative refinement.

Inventive Principle:

Principle #40Composite materials

4Manufacturing precision

If comprehensive optimization processes are implemented to improve simulation accuracy, then device complexity increases, but ease of operation deteriorates

Engineering Contradiction:

Improveaccuracy of fabric drape simulationVSAvoidsimplicity of parameter estimation process

Core Design Contradiction:

Manufacturing precisionVSEase of operation

Solution Approach 1:

The patent implements a self-service system where the neural network automatically performs parameter estimation from fabric images without requiring manual input of physical property parameters. The system autonomously processes images, generates initial parameter estimates, and refines them through simulation feedback, eliminating complex manual操作流程.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20260080626A1Method and apparatus for optimizing garment simulation parameters

Publication Date: 2026.03.19 CLO VIRTUAL FASHION INC

AI summary

A method and device for estimating physical property parameters are disclosed. The method of estimating physical property parameters for a drape simulation of a virtual fabric includes generating a mesh by applying physical property parameters corresponding to the virtual fabric to a neural network, obtaining, based on the mesh, drape data corresponding to a type of target data related to a drape of the virtual fabric, and updating the physical property parameters based on an error between the obtained drape data and the target data.