Advertisement

A cooperative algorithm for the dynamic stereo problem

  • Gerald Leonard Gordon
Track 2: Artificial Intelligence
Part of the Lecture Notes in Computer Science book series (LNCS, volume 507)

Abstract

A basic problem in the study of visual motion is determining temporal correspondence, which is commonly called tracking. The psychophysics community suggests that motion and stereo are cooperative processes, each aiding the other in deciding the final correspondences. In this article, we introduce an algorithm which uses a cooperative process between stereo and motion to attack the dynamic stereo problem. The algorithm, after isolating the objects to be tracking in each frame, lists the possible stereo matches in the left and right views which satisfy certain constraints. It then relies on a general smoothness of motion criterion using (1) the center of gravity of each object in each monocular view, (2) the sizes of the objects, as well as (3) the stereo disparities between possible matched stereo objects. Experiments using tennis balls shows the algorithm to be very promising.

Keywords

Stereo Match Tennis Ball Cooperative Process Actual Trajectory Left View 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    S.N. Antis, Phi motion as a subtraction process, Vision Research, Vol 10, pp.1411–1430, 1970.CrossRefGoogle Scholar
  2. [2]
    G.L. Gordon, On the tracking of featureless objects with occlusion, Proceedings of the IEEE Workshop on Visual Motion, pp. 13–20, Irvine, Ca., 1989.Google Scholar
  3. [3]
    M. Jenkins and J.K. Tsotsos, Applying temporal constraints to the dynamic stereo problem, Computer Vision, Graphics, and Image Processing, Vol. 33, pp. 16–32, 1986.CrossRefGoogle Scholar
  4. [4]
    P.A. Kolers, Aspects of Motion Perception, Pergamon Press, Oxford,1972.Google Scholar
  5. [5]
    M.K. Leung, A.N. Choudhary, J.H. Patel, T.S. Huang, Point matching in a time sequence of stereo image pairs and its parallel implementation on a multiprocessor” Proceedings of the IEEE Workshop on Visual Motion, pp. 321–27, Irvine, Ca., 1989.Google Scholar
  6. [6]
    A. Mitchie, A computation approach to the fusion of stereopsis and kineopsis, Motion Understanding: Robot and Human Vision, ed. by W.N. Martin and J.K. Aggarwal, pp. 81–99, Kluwer Acadmeic Publishers, Boston, 1988.Google Scholar
  7. [7]
    Ramachandran, Invited Address, Proceedings of the IEEE Workshop on Visual Motion, Irvine, Ca., 1989.Google Scholar
  8. [8]
    A. Ratleff, A Study of visual movements determined by form, color, or brightness, Acta Physchol, Vol. 12, pp. 64–70, 1956.Google Scholar
  9. [9]
    I.K. Sethi and R. Jain, Finding trajectories of feature points in a monocular image sequence, IEEE Tran. on PAMI, Vol. 9, pp.56–73, 1987.Google Scholar
  10. [10]
    H. Shariat and K. Price, Motion Estimation using more than two images, Motion Understaning: Robot and Human Vision, ed. by W.N. Martin and J.K. Aggarwal, pp. 143–188, Kluwer Academic Publishers, Boston, 1988.Google Scholar
  11. [11]
    Y. Yasumoto and G. Medioni, Experiments in estimation of 3-dimensional motion parameters from a sequence of image frames, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, San Francisco, pp. 89–94, 1985.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • Gerald Leonard Gordon
    • 1
  1. 1.Department of Computer ScienceDepaul UniversityChicago

Personalised recommendations