Creating custom fit of garments is traditionally a laborious task and requires years of practice for pattern makers to master. Commercial software for pattern drafting and garment simulation emerges as a result and greatly shortens the apparel product development cycle and reduces material cost. The tools provided in these software, however, solve only part of the problem – based on human models that represent target wearers or group of target customers, pattern makers still need trial and error to obtain satisfactory fits in virtual environments. The need to take accurate body measurement and transform the measurements to the 2D patterns; they then test the fitting in the 3D to see if garment features and corresponding body features are well aligned and see if enough ease has been appl...[ Read more ]

Creating custom fit of garments is traditionally a laborious task and requires years of practice for pattern makers to master. Commercial software for pattern drafting and garment simulation emerges as a result and greatly shortens the apparel product development cycle and reduces material cost. The tools provided in these software, however, solve only part of the problem – based on human models that represent target wearers or group of target customers, pattern makers still need trial and error to obtain satisfactory fits in virtual environments. The need to take accurate body measurement and transform the measurements to the 2D patterns; they then test the fitting in the 3D to see if garment features and corresponding body features are well aligned and see if enough ease has been applied to ensure wearing comfort. If the fit in the 3D is not satisfactory, they need to adjust the 2D patterns and test them in 3D again. This tedious and complex approach deters the recruitment of new generation pattern makers. Further, this time-consuming customization process fails to support today’s high demand from online fashion retailing and tailor-made services. Garment modeling and fitting techniques for customization is thus an important research topic.

To overcome the challenges in garment fitting, it is necessary to use 3D simulation techniques to visualize design outcomes and to evaluate the wearing comfort of the garments on human model which has body measurements representative of the target customer. This study contributes to provide a general and flexible framework to construct garment models for custom fit and at the same time ensures that the resulting 2D pattern models are directly linked to garment fabrications and further style developments.

The 2D patterns considered in this study are basic block patterns which are used to cater for fitting and are thus the foundation of garments of all different styles. In our approach, the first step is to quantify the misfit in the 3D garment by a misfit energy and then check the misfit energy associated with the initial 3D garment model draped on the human model. A 2D-to-3D sensitivity matrix is used to linkup 2D pattern model adjustments and the misfit energy in the 3D garment model. As the 2D pattern model adjustments have to conform to the morphological structure of the basic block pattern, we propose also a set of construction constraints that conforms to professional pattern making practice to ensure correct structure of the 2D pattern models.

The 2D pattern models are optimized by gradient descent with a novel step size calculation method; the adjusted 3D garment model is then again evaluated against the 3D human model for fitting. This iterative procedure continues until the garment meets the 3D misfit energy gradient threshold. The algorithm of this fitting process is presented in chapter 4 of the thesis.

To further enhance the wearing comfort, an in-plane strain misfit energy in the 3D garment model is evaluated to ensure that it has minimal strains. The strains due to gravity is decoupled from the overall strains so that the in-plane strain misfit energy reflects only the effect of tightness. The in-plane strain misfit energy combines with the misfit energy in chapter 4 to form the total misfit energy. A novel and efficient in-plane strain misfit energy gradient approximation technique is proposed so that the 2D patterns are adjusted and optimized to reduce the 3D garment’s total misfit energy efficiently. This algorithm is presented in chapter 5 of the thesis.

The simulation results and fabricated garment results are presented in chapter 6 to demonstrate the effectiveness of the proposed methodology in meeting the fitting requirements.

While a lot of work contribute to interactive garment editing and automatic grading, this work provides general and flexible automated fit customization to cater for individual differences and preferences.