ABSTRACT
Capturing 3D human movements that can be used for bio-mechanical analysis requires the sensor systems to be able to follow the motion trajectory of the performer. The operating environment may be highly cluttered. The non-intrusive factor is also very important so that it can be applied in sports competitions. This implies that existing commercial motion capture (mocap) methods that require the subject to wear sensors are not suitable, hence constraining us to just video cameras. We may also have to face the situations when minimal cameras can be used due to constriction and limited resources. This prompted us to draw expertise from both computer vision and computer graphics techniques to accomplish our task. The computer graphics area will provide the knowledge for 3D modeling, animation and synthesis, while the computer vision area provides the learning and analysis capabilities.
