Markerless motion capture of complex full-body movement for character animation

  • Authors:

  • Andrew J. Davison;Jonathan Deutscher;Ian D. Reid

  • Affiliations:

  • Robotics Research Group, Department of Engineering Science, University of Oxford, Oxford OX1 3PJ, UK;Robotics Research Group, Department of Engineering Science, University of Oxford, Oxford OX1 3PJ, UK;Robotics Research Group, Department of Engineering Science, University of Oxford, Oxford OX1 3PJ, UK

  • Venue:
  • Proceedings of the Eurographic workshop on Computer animation and simulation
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

Vision-based full-body tracking aims to reproduce the performance of current commercial marker-based motion capture methods in a system which can be run using conventional cameras and without the use of special apparel or other equipment, improving usability in existing application domains and opening up new possibilities since the methods can be applied to image sequences acquired from any source. We present results from a system able to perform robust visual tracking with an articulated body model, using data from multiple cameras. Our approach to searching through the high-dimensional model configuration space is an algorithm called annealed particle filtering which finds the best fit to image data via multiple-layer propagation of a stochastic particle set. This algorithm efficiently searches the configuration space without the need for restrictive dynamical models, permitting tracking of agile, varied movement. The data acquired can readily be applied to the animation of CG characters. Movie files illustrating the results in this paper may be obtained from http://www.robots.ox.ac.uk/~ajd/HMC/