PARAMETRIC MODELING OF THE HUMAN BODY SHAPE BY STATISTICAL MODEL

This paper proposes a novel method for semantic parametric design of 3D human body. The body shape and their semantic parameters are statistically correlated as a linear system of equations based on existing reference models. Principal component analysis (PCA) is applied to reduce the computational complex while preserve a good modeling capability in the correlation. A convex hull algorithm is developed to quickly determine whether the parameter values input by the user for new model construction is reasonable or not. Test examples demonstrate that this work provides a simple but effective approach to freeform design of human-centric products.