A hybrid approach towards fully automatic 3D marker tracking

ABSTRACT

Motion Capture is a powerful approach to track 3D position, usually utilizing markers. Especially passive markers do not hinder natural motion. Unfortunately, such systems do not provide any information about which anatomical landmark their markers belong to. Multiple manual actions are often required to guide the tracking process. This work presents a hybrid approach for nearly fully automatic identification and tracking of such markers. It encompasses three methods for identification, using PCA-based alignment or tree-based optimization, and tracking, using a neural network with self-organizing characteristics.

REFERENCES

• 1: Dorfmüller-Ulhaas, K. 2003. Robust Optical User Motion Tracking Using a Kalman Filter. Tech. Rep. 2003--6, University of Augsburg, Institut für Informatik.
• 2: Alexander Hornung , Sandip Sar-Dessai, Self-Calibrating Optical Motion Tracking for Articulated Bodies, Proceedings of the 2005 IEEE Conference 2005 on Virtual Reality, p.75-82, March 12-16, 2005 [doi>10.1109/VR.2005.67]
• 3: O'Brien, J. F., Bodenheimer, R., Brostow, G., and Hodgins, J. K. 2000. Automatic Joint Parameter Estimation from Magnetic Motion Capture Data. In Graphics Interface, 53--60.
• 4: Weber, M., Ben Amor, H., and Alexander, T. 2008. Enhancing Motion Capture Performance by Means of an Internal Anthropometric Skeleton Model. In SAE Digital Human Modeling Conference.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.6 Learning
          Subjects: Connectionism and neural nets

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.10 Vision and Scene Understanding
            Subjects: Motion

Keywords:
hybrid tracking, motion capture, neural networks