Software Development for Computer Aided Design and Manufacturing Applications

Based on body scan data and geometric computing methods, this series of studies developed an expert system for fashion design and clothing pattern.
Research includes:

  1. 3D CAD fashion design software－
    Today’s fashion industry uses 2D pattern design software, and no research has been conducted on customizing designs based on human body feature measurements obtained through body scanning. Our 3D CAD fashion design system allows for the draping creation of individualized manikins, mediated by parametric expert data to create basal designs. These designs can be further customized in a 3D editing interface with different settings for cutting, color and pattern, followed by fitting analysis. We expect that this software will significantly reduce garment industry processing hours, rewrite the training content of school programs for future fashion designers and have a major impact on the garment industry.
  2. Pattern Generation and Manufacture from 3D Clothing design－
    Orthogonal geometry transformation methods are used to convert a 3D clothing image into 2D patterns. Formulas based on the apparel expertise of fashion experts are combined with binding constraints on settings for pattern generation. This study proposes post-processing optimization methods for automated surface flattening, using pre-designed surfaces and post-processing appearance results to create a higher degree of consistency with the original 3D design surface. After segmenting the 3D design surface to triangular meshes and flattening them, we then use optimization theory to analyze the degree of distortion arising from expansion from 3D to 2D, along with total deformation energy to assess energy strengths and weaknesses inside the 2D pattern. This distribution is then used as a basis for expanding the resulting split, thus reducing surface energy and optimizing the algorithm adjustment pattern, trying to minimize the total energy generated by the deformation caused by the flattening, establishing several assessment methods, and maintaining user input constraints, etc. Under these constraints, we seek to generate optimal clothing patterns for various individuals. The generated patterns would be fabricated by a large printer to begin actual seam work and fitting. This study was conducted with the cooperation of Professor Li-Juan Wu and Su-Ching Huang of Tainan Women’s College Fashion Design Department.