3D Fabric Drape Analysis for Material Property Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately estimate material property parameters of fabrics due to their flexibility and variability in appearance when worn on a person's body, leading to inconsistent results in fabric design and feel.
Innovation Solution
A method using a trained artificial neural network processes 3D contour shapes of fabrics on geometric objects to estimate material property parameters, incorporating 3D scanning and Gaussian mixture model (GMM) upsampling for training data generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to evaluate fabric properties, then the process is simple, but the measurement precision is insufficient due to fabric flexibility and variability
Solution Approach 1:
The patent replaces traditional mechanical measurement methods with a neural network-based computational system. The neural network processes 3D contour shape data to estimate material property parameters, substituting physical measurement mechanisms with an intelligent computational model that can handle fabric flexibility and variability.
Solution Approach 2:
The patent transitions from 2D fabric patterns to 3D contour shape analysis. By capturing the three-dimensional shape of fabric when draped on a body, the system gains additional spatial dimensions that provide richer information about fabric behavior, enabling more accurate material property estimation while accounting for fabric flexibility.
2Measurement precision
If 3D contour shape processing is used, then material property estimation accuracy improves, but data collection and processing complexity increases
Solution Approach 1:
The patent uses a 3D body scan or digital avatar as a virtual copy of the human body. Instead of requiring complex physical measurements on actual bodies, the system uses this digital representation to simulate and analyze fabric draping, thereby simplifying the data collection process while maintaining measurement accuracy.
Solution Approach 2:
The patent introduces a 3D body scan or digital avatar as an intermediary between the fabric and the measurement system. This intermediary allows the system to capture fabric behavior in a controlled virtual environment, making the detection and measurement process more manageable while preserving the complexity of real-world fabric-body interactions.
3Quantity of substance
If training data is generated using Gaussian mixture model upsampling, then the quantity of training data increases, but the complexity of data generation process increases
Solution Approach 1:
The patent applies Gaussian mixture model upsampling by transforming and replicating existing training data points according to a probability distribution. This parameter-based approach generates additional training data by varying material property parameters within statistical ranges, increasing data quantity while maintaining the underlying data distribution characteristics.
Solution Approach 2:
The data generation process uses the Gaussian mixture model to automatically create training data from existing data patterns. The system serves itself by using the statistical properties of existing data to generate new data points, eliminating the need for manual data collection and reducing the complexity of the data generation process.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Estimating a material property parameter of fabric involves receiving information including a three-dimensional (3D) contour shape of fabric placed over a 3D geometric object, estimating a material property parameter of the fabric used for representing drape shapes of 3D clothes made by the fabric by applying the information to a trained artificial neural network, and providing the material property parameter of the fabric.