Tracking and Classifying of Human Motions with Gaussian Process Annealed Particle Filter

  • Leonid Raskin
  • Michael Rudzsky
  • Ehud Rivlin
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4843)


This paper presents a framework for 3D articulated human body tracking and action classification. The method is based on nonlinear dimensionality reduction of high dimensional data space to low dimensional latent space. Motion of human body is described by concatenation of low dimensional manifolds which characterize different motion types. We introduce a body pose tracker, which uses the learned mapping function from low dimensional latent space to high dimensional body pose space. The trajectories in the latent space provide low dimensional representations of body poses performed during motion. They are used to classify human actions. The approach was checked on HumanEva dataset as well as on our own one. The results and the comparison to other methods are presented.


Latent Space Tracking Algorithm Data Space Motion Type Latent Variable Model 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Christoudias, C.M., Darrell, T.: On modelling nonlinear shape-and-texture appearance manifolds. In: Proc. CVPR, vol. 2, pp. 1067–1074 (2005)Google Scholar
  2. 2.
    Elgammal, A., Lee, C.: Inferring 3d body pose from silhouettes using activity manifold learning. In: Proc. CVPR, vol. 2, pp. 681–688 (2004)Google Scholar
  3. 3.
    Lawrence, N.: Gaussian process latent variable models for visualization of high dimensional data. In: NIPS. Information Processing Systems, vol. 16, pp. 329–336 (2004)Google Scholar
  4. 4.
    Wang, J., Fleet, D., Hetzmann, A.: Gaussian process dynamical models. In: NIPS. Information Processing Systems, pp. 1441–1448 (2005)Google Scholar
  5. 5.
    Deutscher, J., Blake, A., Reid, I.: Articulated body motion capture by annealed particle filtering. In: Proc. CVPR, pp. 2126–2133 (2000)Google Scholar
  6. 6.
    Alt, H., Knauer, C., Wenk, C.: Matching polygonal curves with respect to the frèchet distance. In: Ferreira, A., Reichel, H. (eds.) STACS 2001. LNCS, vol. 2010, pp. 63–74. Springer, Heidelberg (2001)CrossRefGoogle Scholar
  7. 7.
    Isard, M., Blake, A.: Condensation - conditional density propagation for visual tracking. International Journal of Computer Vision 29(1), 5–28 (1998)CrossRefGoogle Scholar
  8. 8.
    Sidenbladh, H., Black, M., Fleet, D.: Stochastic tracking of 3d human figures using 2d image motion. In: Vernon, D. (ed.) ECCV 2000. LNCS, vol. 1842, pp. 702–718. Springer, Heidelberg (2000)CrossRefGoogle Scholar
  9. 9.
    Mikolajczyk, K., Schmid, K., Zisserman, A.: Human detection based on a probabilistic assembly of robust part detectors. In: Proc. ECCV, vol. 1, pp. 69–82 (2003)Google Scholar
  10. 10.
    Rohr, K.: Human movement analysis based on explicit motion models. Motion-Based Recognition 8, 171–198 (1997)Google Scholar
  11. 11.
    Wang, Q., Xu, G., Ai, H.: Learning object intrinsic structure for robust visual tracking. In: Proc. CVPR, vol. 2, pp. 227–233 (2003)Google Scholar
  12. 12.
    Tenenbaum, J., de Silva, V.: A global geometric framework for nonlinear dimensionality reduction. Science 290, 2319–2323 (2000)CrossRefGoogle Scholar
  13. 13.
    Roweis, S., Saul, L.: Nonlinear dimensionality reduction by locally linear embedding. Science 290, 2323–2326 (2000)CrossRefGoogle Scholar
  14. 14.
    Urtasun, R., Fleet, D., Fua, P.: 3d people tracking with gaussian process dynamical models. In: Proc. CVPR, vol. 1, pp. 238–245 (2006)Google Scholar
  15. 15.
    Sato, K., Aggarwal, J.: Recognizing two-person interactions in outdoor image sequences. In: IEEE Workshop on Multi-Object Tracking, IEEE Computer Society Press, Los Alamitos (2001)Google Scholar
  16. 16.
    Park, S., Aggrawal, J.: Recognition of human interactions using multiple features in a grayscale images. In: Proc. ICPR, vol. 1, pp. 51–54 (2000)Google Scholar
  17. 17.
    Hongeng, S., Bremond, F., Nevatia, R.: Representation and optimal recognition of human activities. In: Proc. CVPR, vol. 1, pp. 818–825 (2000)Google Scholar
  18. 18.
    Park, J., Park, S., Aggrawal, J.: Video retrieval of human interactions using model-based motion tracking and multi-layer finite state automata. In: Bakker, E.M., Lew, M.S., Huang, T.S., Sebe, N., Zhou, X.S. (eds.) CIVR 2003. LNCS, vol. 2728, Springer, Heidelberg (2003)Google Scholar
  19. 19.
    Deutscher, J., Reid, I.: Articulated body motion capture by stochastic search. International Journal of Computer Vision 61(2), 185–205 (2004)CrossRefGoogle Scholar
  20. 20.
    Balan, A., Sigal, L., Black, M.: A quantitative evaluation of video-based 3d person tracking. In: VS-PETS. IEEE Workshop on Visual Surveillance and Performance Evaluation of Tracking and Surveillance, pp. 349–356. IEEE Computer Society Press, Los Alamitos (2005)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  • Leonid Raskin
    • 1
  • Michael Rudzsky
    • 1
  • Ehud Rivlin
    • 1
  1. 1.Computer Science Department, Technion—Israel Institute of Technology, Haifa, 32000Israel

Personalised recommendations