Learning flexible models from image sequences

  • Adam Baumberg
  • David Hogg
Shape Modelling
Part of the Lecture Notes in Computer Science book series (LNCS, volume 800)


The “Point Distribution Model”, derived by analysing the modes of variation of a set of training examples, can be a useful tool in machine vision. One of the drawbacks of this approach to date is that the training data is acquired with human intervention where fixed points must be selected by eye from example images. A method is described for generating a similar flexible shape model automatically from real image data. A cubic B-spline is used as the shape vector for training the model. Large training sets are used to generate a robust model of the human profile for use in the labelling and tracking of pedestrians in real-world scenes. Furthermore, an extended model is described which incorporates direction of motion, allowing the extrapolation of direction from shape.


Control Point Background Image Live Video Active Shape Model Shape Vector 
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.


  1. 1.
    Kass M., Witkin A., and Terzopoulos D. Snakes: Active contour models. First International Conference on Computer Vision, pages 259–268, 1987.Google Scholar
  2. 2.
    Yuille A.L., Cohen D.S., and Hallinan P. Feature extraction from faces using deformable templates. Computer Vision and Pattern Recognition, pages 104–109, 1989.Google Scholar
  3. 3.
    Blake A., Curwen R., and Zisserman A. A framework for spatio-temporal control in the tracking of visual contours. International Journal of computer Vision, 1993.Google Scholar
  4. 4.
    Cootes T.J., Taylor C.J., Cooper D.H., and Graham J. Training models of shape from sets of examples. In British Machine Vision Conference, pages 9–18, September 1992.Google Scholar
  5. 5.
    Hogg D. Model-based vision: A program to see a walking person. Image and Vision Computing, 1(1):5–20, 1983.Google Scholar
  6. 6.
    Rohr K. Incremental recognition of pedestrians from image sequences. Computer Vision and Pattern Recognition, pages 8–13, 1993.Google Scholar
  7. 7.
    Pentland A. and Horowitz B. Recovery of non-rigid motion and structure. IEEE Trans. on Pattern Analysis and Machine Intelligence, 13(7):730–742, July 1991.Google Scholar
  8. 8.
    Murphy N., Byrne N., and O'Leary K. Long sequence analysis of human motion using eigenvector decomposition. In Proc. SPIE, September 1993.Google Scholar
  9. 9.
    Worrall A. and Hyde J. A fast algorithm for background generation. VIEWS Working Paper RU-03-WP-T. Scholar
  10. 10.
    Hill A., Thornham A., and Taylor C.J. Model-based interpretation of 3d medical images. In British Machine Vision Conference, volume 2, pages 339–349, 1993.Google Scholar
  11. 11.
    Bartels R., Beatty J., and Barsky B. An Introduction to Splines for use in Computer Graphics and Geomteric Modeling. Morgan Kaufmann, 1987.Google Scholar
  12. 12.
    Li-Qun X., Young D., and Hogg D. Building a model of a road junction using moving vehicle information. In British Machine Vision Conference, pages 443–452, September 1992.Google Scholar
  13. 13.
    Cootes T.F. and Taylor C.J. Active shape models — 'smart snakes'. In British Machine Vision Conference, pages 276–285, September 1992.Google Scholar
  14. 14.
    Shapiro L. and Brady M. Rejecting outliers and estimating errors in an orthogonal regression framework. Ouel, Robotics Research Group, University of Oxford, February 1993.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1994

Authors and Affiliations

  • Adam Baumberg
    • 1
  • David Hogg
    • 1
  1. 1.School of Computer StudiesUniversity of LeedsLeedsEngland

Personalised recommendations