A Novel Approach to Classify and Characterize Body Type of Chinese Female Adults by Using Principal Component Analysis and Dynamic Cluster Algorithm

摘要

New developments of computer science and information technology make it possible to realize mass customization in apparel industry. These years many scholars and researchers have paid attention to classify and characterize human body shape to accelerate the process of making paper patterns. This paper based on an anthropometric survey of 280 Chinese women aged from 18 to 50 by [TC] 2 non-contact 3D body scanning system. By means of principal component analysis (PCA), 39 measurement items were transformed into 7 uncorrelated principal factors. Furthermore, these principal factors were given professional definitions according to their eigenvectors. In order to make effective criteria for classifying female body type, all samples are sorted by these factors using dynamic samples cluster algorithm. In conclusion, it provides a new way to study female body type and will be useful to further somatotype research and practical garment manufacturing for mass customization.