Estimating human body characteristics under clothing using a statistical body shape model
Authors: Byoung-Keon Park, B. D. Corner, M. Kearney, Matthew P. Reed
Whole-body surface scanning technology has revolutionized anthropometry, providing an efficient way to capture detailed body characteristics. Scanning protocols generally require minimally-clad subjects, thus obtaining a body model from a scan of a person with normal clothing still remains a challenge. The current study aims to estimate body shape under clothing or equipment using a statistical body shape model (SBSM). The SBSM was fitted to 100 soldier scans with four different levels of clothing and gear to estimate the body shapes using an inscribed-fitting method. The inscribed-fitting method finds the body shape by finding the maximum body volume available in the body shape space of the model. A rapid fitting method developed in our previous work was used to search a body shape in the body shape space. The result showed the good accuracy in estimating body shape under clothing and gear, with the mean error of 11.2 mm compared to the actual body shape.