Three-dimensional Technology for Apparel Mass Customization: Part II: Human Body Modeling from Unorganized Range Data

Abstract

This paper presents a human body modeling method using the B-spline surface approximation algorithm and unorganized range data from 3D scanning systems. The scan data are firstly reorganized to have a uniform density through the Delaunay triangulation and linear interpolation. The scanned body is divided into six cylindrical segments, and the cubic B-spline surface approximation is performed on each segment separately. The individual B-spline surfaces are then merged to create a seamless model, which not only dramatically reduces the size of the data needed to reconstruct the surface but also smooths and completes the scan data. The fitted model is essential for e-tape measurement of the body and virtual try-on of clothes.