Personalized Clothing Pattern Generation Driven by Body Shape Data

Addressing the inefficiencies of existing pattern- making techniques and their inability to meet personalized demands, this paper investigates an automated method for generating personalized garment patterns based on human body size information. Utilizing methods such as LASSO regression, spline functions, and neural networks, we establish a model capable of automatically extracting key features from human data and pattern templates to produce size-adaptable garment patterns. Specifically, multi-factor LASSO regression is employed to extract significant features from body measurements, spline functions are used to fit garment pattern lines and extract garment feature points, and finally, neural networks are applied to integrate human and garment features to obtain adaptively scaled fashion patterns. Through ablation experiments, the model's high predictive accuracy is demonstrated. The implementation of this project will provide a new technological approach to achieving automation and personalization in the garment industry.