A Pose Invariant Statistical Shape Model for Human Bodies

Abstract

We present a complete pipeline for constructing a statistical shape model that is invariant to deviations in the scan pose while encoding the space of human pose and body shape in an efficient manner. A dense cross-parameterization between a large set of high-quality 3D scans is computed using a fast and robust volume aware non-rigid registration method. Our approach uses a novel encoding that automatically decorrelates shape and pose leading to a statistical model that is oblivious under transformations induced by pose. This allows us to efficiently compensate pose variations in captured input data leading to a compact representation for pose as well as body shape. We present a local as well as a global skeletal encoding and compare both approaches. Finally, we analyze the generalization properties and accuracy of our approach against two state-of-the-art methods. We apply our model to the data clustering problem and use it as a prior for non-rigid shape matching.

Volume Extraction from Body Scans for Bra Sizing

“…For extracting a mesh representation we use a two-step approach. In the first step, we register a coarse statistical model to the point cloud (see [10]). To further refine the mesh, non-rigid registration based on [6,11] is performed.…”

Measuring the efficacy of positioning aids for capturing 3D data in different clothing configurations and postures with a high-resolution whole-body scanner

“…In research, different approaches have been taken to overcome this deficiency. Ballester et al [15] and Colaianni et al [16] presented software-supported approaches In contrast to software-supported posture harmonisation, mechanical positioning aids potentially reduce postural variability of both semi-nude and clothed participants. In earlier research only a few attempts have been made to counter human variability, i.e.…”

“…Subsequently, scan two and three were superimposed and horizontal sections taken through seven of the previously extracted anatomical landmarks in the standing posture and six in the seated posture (Figure 7). [16] Contrary to earlier research [19], the software package PolyWorks (InnovMetric) was used instead of Anthroscan for taking and gauging horizontal sections. Compared to Anthroscan, PolyWorks has the advantage that different colours can be assigned to the sections of the individual scans (Figure 8b).…”

Aerodynamic effects and performance improvements of running in drafting formations

“…The geometry represents an average of 59 individual track athletes, scanned in-house and positioned into a running pose. The obtained poseinvariant statistical model for the female human body was reconstructed from the high-quality scan data using the method outlined by Colaianni et al (2014). The axial and lateral spacing between the runners is chosen according to estimates of realistic running conditions.…”