Page 199 - Anthropometry, Apparel Sizing and Design

192 Anthropometry, Apparel Sizing and Design

7.4 Dynamic recording of body surfaces

So far, movement has been digitalized with motion capture technology (Watkins and
Dunne, 2015). Defined points are tracked on the body or on the silhouette and turned
into data that a computer can read. Animation and control of applications such as
video games and computer-animated films or biomechanical analyses in medicine
and sports science are among the purposes.
The output is an animated, sometimes very abstracted skeleton, as the focus in bio-
mechanics lies on the range of movement of individual joints. Information about
changes in body surface area cannot yet be reliably derived using motion capture tech-
nology, although scientific research has already been carried out in this area. Dynamic
recording of the whole 3-D surface using motion capture technology was presented by
Park and Hodgins (2006). With 350 markers the geometry of the skin and its deforma-
tion was only recorded limitedly but was then transferred to a static 3-D scan and inter-
polated. This achieves a higher resolution of the surface mesh but still not an exact
representation accurate of the skin surface. For a better resolution, more markers have
to be used, which is limited for the currently available motion capture technology. The
reflective markers are relatively large, and the effort of calibration grows with each
marker. The automatic tracking when there are too many markers also has to be com-
plexly corrected by hand. Meanwhile, there are further research approaches. However,
the majority of studies work with interpolation of points and are also coupled with great
effort (Aguiar, 2014; Wu et al., 2011; Rose and Murphy, 2015).
Four-dimensional technologies have been mostly used in medicine so far. By
means of computed tomography, magnetic resonance tomography, or ultrasound,
the interior of the body (e.g., organs, skeleton, and embryos) is visualized in three
dimensions and in stop motion (Li et al., 2008; Rivera-Rivera et al., 2015;
Schreibmann et al., 2014; Smith et al., 2005). In recent years, scanner systems have
also been developed to record the body surface that are able to display areas of the
body faster and more precisely, such as the face, feet, or back, in both three-
dimensional and motion (Vialux, n.d.; 3dMD, 2019; Lane, 2013; kle-point, n.d.;
Sinfomed, n.d.). They mainly work with structured light. The three-dimensional
recording by means of structured light enables the digitization of recorded surfaces
in the form of point clouds. Here, projectors project a defined pattern (e.g., stripes
and checks) onto the area to be detected, which is captured by cameras from defined
angular positions. Using various algorithms and measuring principles, such as the tri-
angulation method, light-section method, coded light approach, and phase shift
method, the detected areas can then be reconstructed in three dimensions. On this
basis, series recordings allow the scanning of movement. The research group led
by Michael Black at the Max Planck Institute for Intelligent Systems (MPI-IS) works,
according to its own statement, with the world’s first 4-D body scanner. This is based
on a system of 3dMD with 22 stereo cameras, 22 color cameras, and spot pattern pro-
jectors. The system enables movement recording with 60 pictures per second (pps).
With good image quality, it can be assumed that the more pps made, the better and
more detailed the movement can be detected. The scanner was individually developed
for the MPI-IS. The challenge with four-dimensional recording is that multiple sensors