Error in representation of standard anthropometric data by human figure models

A geometric human figure model instance is designed to resemble a human counterpart, and to represent some of its known properties. The geometrical representation of standard anthropometric data by a $3D$ human figure almost always depends on the use of approximations to establish numerical constraints that assure consistency in the definition of selected pairs of landmark and joint center locations in the figure. Further, the geometric representation of such numerically constrained data may suffer from error, because of the use of simplistic (from an anthropometric standpoint) geometric models that do not offer the capabilities required to support such constraints, which may result in a need for the consideration of additional approximations in order to complete the model. Thus, there are at least two major sources of error in the anthropometric data representation process: the approximations used to establish the numerical constraints, and the approximations used when representing such constraints with the geometric model. This work provides a formal definition of this representation error, in the context of the representation process formulated as a series of transformations. Also, this work includes suggested multivariate criteria and techniques for assessment and exploration of this error, and provides guide-lines for creating a software tool that implements those criteria and techniques. To our knowledge, our approach is novel. We do not know of any state-of-the-art figure model currently available that considers this representation error explicitly, or provides formally stated criteria and/or automated computer software capabilities for its assessment and exploration. The main contribution of this research is to establish a foundation for the explicit consideration of the anthropometric representation error. Such foundation should benefit both the users and the designers of figure models representing anthropometric data.