New human body shape descriptor based on anthropometrics points

This paper will focus on the issue of human body dissimilarity detection from 3D bodyscan. A new 3D human body shape descriptor is proposed as well as a global geometric shape analysis of body shape surfaces coupled with anthropometrics points. The aim of this research is then to establish a new methodology of human body morphology shapes detection in order to define the morphotypes of a given population. A computation of the geodesic distributions based on anthropometrics feature points for human torso provides quantitative information about their similarities. The Euclidean distance is the metric used for the comparison of the shape descriptors. The k-means clustering technique is then implemented to define the most relevant morphologies. Our methodology is then evaluated on 3D scan database of 53 female. The study may be attracted for further researchers from several research communities including pattern recognition, computer graphics, computer vision, anthropometry, human morphology, data analysis and mass customization.