Qualitative Characterization and Use of Prior Information

  • Martin Eriksson
  • Stefan Carlsson
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2749)

Abstract

The use of prior information by learning from training data is used increasingly in image analysis and computer vision. The high dimensionality of the parameter spaces and the complexity of the probability distributions however often makes the exact learning of priors an impossible problem, requiring an excessive amount of training data that is seldom realizable in practise. In this paper we propose a weaker form of prior estimation which tries to learn the boundaries of impossible events from examples. This is equivalent to estimating the support of the prior distribution or the manifold of possible events. The idea is to model the set of possible events by algebraic inequalities. Learning proceeds by selecting those inequalities that show a consistent sign when applied to the training data set. The manifold of possible events estimated in this way will in general represent the qualitative properties of the events. We give example of this in the problems of restoration of handwritten characters and automatically tracked body locations

Keywords

Bayesian Inference Prior Information Human Motion Consistent Sign Handwritten Character 
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.

References

  1. 1.
    O. Aichholzer, F. Aurenhammer, and H. Krasser. Enumerating order types for small point sets with applications. In Proc. of the 17:th Ann. ACM Symp. Computational Geometry, pages 11–18, 2001.Google Scholar
  2. 2.
    O. Aichholzer, F. Aurenhammer, and H. Krasser. Points and combinatorics. Special Issue on Foundations of Information Processing of TELEMATIK, 1(7):12–17, 2002.Google Scholar
  3. 3.
    R. Bowden. Learning statistical models of human motion. In Proc. of the IEEE Workshop on Human Modeling, Analysis and Synthesis, CVPR, 2000.Google Scholar
  4. 4.
    M. Brand. Shadow puppetry. In Proceedings of the International Conference on Computer Vision, pages 1237–1244, 1999.Google Scholar
  5. 5.
    C. Bregler. Learning and recognizing human dynamics in video-sequences. In Proc. of the IEEE International Conference on Computer Vision and Pattern Recognition, pages 568–574, 1997.Google Scholar
  6. 6.
    Jos Gomes and Aleksandra Mojsilovic. A variational approach to recovering a manifold from sample points. In Proc. of the European Conference on Computer Vision, pages 3–17, 2002.Google Scholar
  7. 7.
    Michael E. Leventon and William T. Freeman. Bayesian estimation of 3d human motion from an image sequence. Technical Report TR-98-06, MERL-A Mitsubishi Research Laboratory, 1998.Google Scholar
  8. 8.
    Vladimir Pavlovic, James M. Rehg, Tat-Jen Cham, and Kevin P. Murphy. A dynamic bayesian network approach to figure tracking using learned dynamic models. In Proc. of the IEEE International Conference on Computer Vision, volume 1, pages 94–101, 1999.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Martin Eriksson
    • 1
  • Stefan Carlsson
    • 1
  1. 1.Dept. of Numerical Analysis and Computing ScienceRoyal Institute of Technology, (KTH)StockholmSweden

Personalised recommendations