Modeling body shape from surface landmark configurations

Abstract

Detailed statistical models of body size and shape are valuable for wide range of statistical analyses. Most body shape models represent a single posture, usually standing. Previous efforts to model both posture and body shape have parameterized posture using joint angles. This paper presents a statistical model of body shape in supported seated postures using a posture measured derived from surface landmarks rather than internal joint locations and angles. This method is not limited by a particular kinematic linkage deformation and so is particularly well suited to model the effects on body shape of posture changes in complex linkages such as the spine or shoulder. © 2013 Springer-Verlag.