PerfectTailor: Scale-Preserving 2D Pattern Adjustment Driven by 3D Garment Editing

Computer Graphics and Applications - Journal paper -
Presented at Geometric Modeling and Processing (GMP) 2025

Anran Qi and Takeo Igarashi
The University of Tokyo
Paper Presentatioin Slides

Motivation

Traditional garment editing workflow: Edit 2D pattern, then run simulation to see 3D, as shown in the video below.

Pattern: ©CLO Brasil

But this workflow has two problems:

  • The designer must mentally convert target 3D shape to 2D pattern.
  • It requires expertise, which is difficult for non-experts

Our goal is to enable non-experts to directly edit garments in 3D and automatically obtain the corresponding 2D patterns. Previous approaches with similar goals rely on 'flattening' the 3D garment using geometric surface parameterization. But standard flattening methods fail to consider domain specific constraints. More specifically, they

Motivation Image

Method

Image 1 Image 2

Core idea1: we propose to memorize the local scale difference between the 2D pattern and the 3D drape of the initial design and use it when updating the 2-D pattern to accommodate user editing on the 3-D drape.

Core idea2: we propose an as-original-as-possible constraint to preserve the discrete tangent of boundary vertices.

User Editing Video

The left is the 3D window displaying the 3D garment on a human body. Red dots indicate key feature points on the surface of the human body (e.g., front neck point). The right is the 2-D pattern window showing the sewing pattern. The top shows the functions of the system. The supported functions include sketch2cut, extend, shorten, tighten, scale and scale by seamline.

BibTeX

@ARTICLE{PerfectTailor2024,
          author={Qi, Anran and Igarashi, Takeo},
          journal={IEEE Computer Graphics and Applications},
          title={PerfectTailor: Scale-Preserving 2-D Pattern Adjustment Driven by 3-D Garment Editing},
          year={2024},
          volume={44},
          number={4},
          pages={126-132},
          doi={10.1109/MCG.2024.3378171}}