3D SSD Tracking from Uncalibrated Video

  • Dana Cobzas
  • Martin Jagersand
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3667)


In registration-based motion tracking precise pose between a reference template and the current image is determined by warping image patches into the template coordinates and matching pixel-wise intensities. Efficient such algorithms are based on relating spatial and temporal derivatives using numerical optimization algorithms. We extend this approach from planar patches into a formulation where the 3D geometry of a scene is both estimated from uncalibrated video and used in the tracking of the same video sequence. Experimentally we compare convergence and accuracy of our uncalibrated 3D tracking to previous approaches. Notably, the 3D algorithm can successfully track over significantly larger pose changes than ones using only planar regions. It also detects occlusions and removes/inserts tracking regions as appropriate in response.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Lowe, D.: Fitting parameterized three-dimensional models to images. PAMI 13, 441–450 (1991)Google Scholar
  2. 2.
    Marchand, E., Bouthemy, P., Chaumette, F.: A 2d-3d model-based approach to real-time visual tracking. IVC 19, 941–955 (2001)Google Scholar
  3. 3.
    Drummond, T., Cipolla, R.: Real-time visual tracking of complex structures. PAMI 24, 932–946 (2002)Google Scholar
  4. 4.
    Armstrong, M., Zisserman, A.: Robust object tracking. In: Second Asian Conference on Computer Vision, pp. 58–62 (1995)Google Scholar
  5. 5.
    Toyama, K., Hager, G.: Incremental focus of attention for robust vision-based tracking. IJCV 35, 45–63 (1999)CrossRefGoogle Scholar
  6. 6.
    Szeliski, R.: Video mosaics for virtual environments. IEEE Computer Graphics and Applications, 22–30 (1996)Google Scholar
  7. 7.
    Lucas, B., Kanade, T.: An iterative image registration technique with an application to stereo vision. In: Int. Joint Conf. on Artificial Intelligence (1981)Google Scholar
  8. 8.
    Hager, G., Belhumeur, P.: Efficient region tracking with parametric models of geometry and illumination. PAMI 20, 1025–1039 (1998)Google Scholar
  9. 9.
    Baker, S., Matthews, I.: Lucas-Kanade 20 Years On: A Unifying Framework. Technical Report CMU-RITR02-16 (2002)Google Scholar
  10. 10.
    Jurie, F., Dhome, M.: Hyperplane approximation for template matching. PAMI 24, 996–1000 (2002)Google Scholar
  11. 11.
    Gleicher, M.: Projective registration with difference decomposition. In: CVPR 1997, pp. 331–337 (1997)Google Scholar
  12. 12.
    Horn, B.: Computer Vision. MIT Press, Cambridge, Mass (1986)Google Scholar
  13. 13.
    Hartley, R.I., Zisserman, A.: Multiple View Geometry in Computer Vision. Cambridge University Press, Cambridge (2000)zbMATHGoogle Scholar
  14. 14.
    Werner, T., Pajdla, T., Urban, M.: Practice of 3d reconstruction from multiple uncalibrated unorganized images. In: Czech Pattern Recognition Workshop (2000)Google Scholar
  15. 15.
    Triggs, W.: Auto-calibration and the absolute quadric. In: CVRP, pp. 609–614 (1997)Google Scholar
  16. 16.
    Cobzas, D., Jagersand, M.: A Comparison of Viewing Geometries for Augmented Reality. In: Bigun, J., Gustavsson, T. (eds.) SCIA 2003. LNCS, vol. 2749, pp. 501–508. Springer, Heidelberg (2003)CrossRefGoogle Scholar
  17. 17.
    line mpeg movies of the experiments are available. See videoX at,

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Dana Cobzas
    • 1
  • Martin Jagersand
    • 2
  1. 1.INRIA Rhone-AplesMontbonnotFrance
  2. 2.Computing ScienceUniversity of AlbertaEdmontonCanada

Personalised recommendations