Advances in Vision-Based Human Body Modeling - 3D Modeling and Animation: Synthesis and Analysis Techniques for the Human Body - page 20

data. Other techniques for anatomical parts recovering or biomedical applica-
tions were presented in Weng, Yang and Pierson (1996) and Tognola et al.
(2002). The first one is based on laser spot and two CCD cameras system to
recover the 3D data, while the second one is based on an optical flow approach
(the object remains stationary while the camera undergoes translational motion).
Barron & Kakadiaris (2000) present a four-step technique for estimating a
human's anthropometric measurements from a single image. Pose and anthro-
pometric measurements are obtained by minimizing a cost function that com-
putes the difference between a set of user-selected image points and the
corresponding projected points of a 3D stick model.
Finally, motion analysis systems, which are based on the study of kinematics
and dynamics parameters, allow detection of movement disabilities of a given
patient. Marzani et al. (1997) and Marzani, Calais & Legrand (2001) present a
system for the analysis of movement disabilities of a human leg during gait. The
proposed system is based on grey-level image processing without the need of
markers. Superquadric surfaces are used to model the legs. This system can be
used in human motion analysis for clinical applications, such as physiotherapy.
Conclusions
Human body modeling is a relatively recent research area with a higher
complexity than the classical rigid object modeling. It takes advantage of most
of the techniques proposed within the rigid object modeling community, together
with a prior-knowledge of human body movements based on a kinematics and
dynamics study of the human body structure. The huge amount of articles
published during the last years involving 3D human body modeling demonstrates
the increasing interest in this topic and its wide range of applications. In spite of
this, many issues are still open (e.g., unconstrained image segmentation, limita-
tions in tracking, development of models including prior knowledge, modeling of
multiple person environments, real-time performance). Each one of these topics
represents a stand-alone problem and their solutions are of interest not only to
human body modeling research, but also to other research fields.
Unconstrained image segmentation remains a challenge to be overcome. An-
other limitation of today's systems is that commonly the motion of a person is
constrained to simple movements with a few occlusions. Occlusions, which
comprise a significant problem yet to be thoroughly solved, may lead to erroneous
tracking. Since existence and accumulation of errors is possible, the systems
must become robust enough to be able to recover any loss of tracking. Similarly,
techniques must be able to automatically self-tune the model's shape param-