Design and Product Development

Leveraging Fashion E-commerce Data and Computer Graphics Toward Automated Pattern Making for Pants: A Preliminary Study

---

Abstract

Ergonomic pattern-making, based on anthropometric measurements, is essential to ensuring fit and comfort but faces challenges including cost and privacy issues. This study explores an economical way of creating patterns for pants by leveraging computer graphics and fashion e-commerce data, such as model and garment sizes, which are readily available online. In this study, 100% cotton denim jeans were selected as sample pants. Computer graphics algorithms were applied to an image of a pair of jeans to automatically detect edges from the image and generate pants patterns by calculating key measurements, such as the waist, hip, crotch, and leg openings. For evaluation, the same pair of jeans was purchased and the patterns were manually traced and digitized into Optitex PDS v.21. The resulting patterns were compared with the algorithm-generated pattern and evaluated using Clo3D v.7. The hip measurements showed the biggest difference between the two patterns, where the algorithm-generated pattern had crooked side seams. In addition, the horizontal lines (e.g., the hip and leg openings) appeared more distorted because those parts were photographed closer to the lens and displayed larger than the vertical lines. While some aspects of geometric distortion in imagery could be improved, this study presented a new method for generating patterns for jeans by synthesizing publicly available e-commerce data and computer graphics. Future studies on deep learning-based models, such as segmentation, could allow for the extraction of more precise edge points for use as critical measurement points.

Keywords: Automated pattern making, pants pattern, computer graphics

How to Cite:

Kong, D. & Baytar, F., (2025) “Leveraging Fashion E-commerce Data and Computer Graphics Toward Automated Pattern Making for Pants: A Preliminary Study”, International Textile and Apparel Association Annual Conference Proceedings 81(1). doi: https://doi.org/10.31274/itaa.17912