Fast and Automatic Identification of Contours for Girth Measurement on 3D Human Models with Variant Postures

Abstract

AbstractWe develop a novel method for fast and automatic identification of characteristic contours for measuring girths on digital 3D human models with various stable postures. For fast identification, the characteristic contours with salient topology and geometric features are selected among the multi-source geodesic iso-contours, which are constructed under a new Morse function as using multi-source points to compute geodesic distance instead of using solely one source point. The resulting iso-contours are stable in a wide range of postures due to the bending invariance of geodesic distance. Based on the characteristic contours, we develop an efficient algorithm to extract the contours for girth measurement. We present a number of experiments on 3D human body samples with variant postures to demonstrate the validity of our approach.