Measuring the Self-Consistency of Stereo Algorithms

  • Yvan G. Leclerc
  • Q.-Tuan Luong
  • P. Fua
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1842)


A new approach to characterizing the performance of point-correspondence algorithms is presented. Instead of relying on any “ground truth’, it uses the self-consistency of the outputs of an algorithm independently applied to different sets of views of a static scene. It allows one to evaluate algorithms for a given class of scenes, as well as to estimate the accuracy of every element of the output of the algorithm for a given set of views. Experiments to demonstrate the usefulness of the methodology are presented.


Window Size Mahalanobis Distance Scatter Diagram Camera Parameter Stereo Match 
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.
    P. Anandan. A computational framework and an algorithm for the measurement of visual motion. IJCV, 2:283–310, 1989.CrossRefGoogle Scholar
  2. 2.
    N. Ayache and F. Lustman. Fast and reliable passive trinocular stereovision. In ICCV, pages 422–427, 1987.Google Scholar
  3. 3.
    K. Cho, P. Meer, and J. Cabrera. Performance assessment through bootstrap. PAMI, 19(11):1185–1198, November 1997.Google Scholar
  4. 4.
    G. Csurka, C. Zeller, Z. Zhang, and O. Faugeras. Characterizing the uncertainty of the fundamental matrix. CVGIP-IU, 1996.Google Scholar
  5. 5.
    O. Faugeras, P. Fua, B. Hotz, R. Ma, L. Robert, M. Thonnat, and Z. Zhang. Quantitative and Qualitative Comparison of some Area and Feature-Based Stereo Algorithms. In W. Forstner and S. Ruwiedel, editors, International Workshop on Robust Computer Vision: Quality of Vision Algorithms, pages 1–26, Karlsruhe, Germany, March 1992.Google Scholar
  6. 6.
    W. Forstner. On the geometric precision of digital correlation. In International archives of photogrammetry and remote sensing, volume 24-III, pages 176–189, Helsinki, 1982.Google Scholar
  7. 7.
    P. Fua. Combining Stereo and Monocular Information to Compute Dense Depth Maps that Pres erve Depth Discontinuities. In Int. Joint Conf. on AI, pages 1292–1298, Sydney, Australia, August 1991.Google Scholar
  8. 8.
    P. Fua and Y. G. Leclerc. Object-Centered Surface Reconstruction: Combining Multi-Image Stereo and Shading. IJCV, 16:35–56, 1995.CrossRefGoogle Scholar
  9. 9.
    T. Kanade and M. Okutomi. A stereo matching algorithm with an adaptive window: Theory and experiment. PAMI, 16(9):920–932, 1994.Google Scholar
  10. 10.
    Y. Leclerc, Q.-T. Luong, and P. Fua. A framework for detecting changes in terrain. In Proc. Image Understanding Workshop, pages 621–630, Monterey, CA, 1998.Google Scholar
  11. 11.
    Y. Leclerc, Q.-T. Luong, and P. Fua. Self-consistency: a novel approach to characterizing the accuracy and reliability of point correspondence algorithms. In Proc. Image Understanding Workshop, pages 793–807, Monterey, CA, 1998.Google Scholar
  12. 12.
    Y. Leclerc, Q.-T. Luong, and P. Fua. Characterizing the performance of multiple-image point-correspondence algorithms using self-consistency. In Proceedings of the Vision Algorithms: Theory and Practice Workshop (ICCV99), Corfu, Greece, September 1999.Google Scholar
  13. 13.
    Y. Leclerc, Q.-T. Luong, and P. Fua. Detecting changes in 3-d shape using self-consistency. In Proc. Conference on Computer Vision and Pattern Recognition, Hilton Head, South Carolina, June 1999.Google Scholar
  14. 14.
    Y. Leclerc, Q.-T. Luong, and P. Fua. Self-consistency: a novel approach to characterizing the accuracy and reliability of point correspondence algorithms. In Proceedings of the One-day Workshop on Performance Characterisation and Bench-marking of Vision Systems, Las Palmas de Gran Canaria, Canary Islands, Spain, 1999.Google Scholar
  15. 15.
    L. Matthies. Stereo vision for planetary rovers: Stochastic modeling to near real-time implementation. IJCV, 8(1):71–91, July 1992.Google Scholar
  16. 16.
    R. Mohr, F. Veillon, and L. Quan. Relative 3d reconstruction using multiple uncalibrated images. In CVPR, pages 543–548, NYC, 1993.Google Scholar
  17. 17.
    W. Førstner. Diagnostics and performance evaluation in computer vision. In Performance versus Methodology in Computer Vision, NSF/ARPA Workshop, Seattle, WA, 1994.Google Scholar
  18. 18.
    R. Szeliski. Prediction error as a quality metric for motion and stereo. In ICCV99, Corfu, Greece, September 1999.Google Scholar
  19. 19.
    R. Szeliski and S.B. Kang. Recovering 3d shape and motion from image streams using nonlinear least squares. JVCIR, pages 10–28, 1994.Google Scholar
  20. 20.
    R. Szeliski and R. Zabih. An experimental comparison of stereo algorithms. In Proceedings of the Vision Algorithms: Theory and Practice Workshop (ICCV99), Corfu, Greece, September 1999.Google Scholar
  21. 21.
    P.H.S. Torr and A. Zissermann. Performance characterization of fundamental matrix estimation under image degradation. mva, 9:321–333, 1997.Google Scholar
  22. 22.
    M. Yachida, Y. Kitamura, and M. Kimachi. Trinocular vision: New approach for correspondence problem. In ICPR, pages 1041–1044, 1986.Google Scholar
  23. 23.
    S. Yi, R.M. Haralick, and L.G. Shapiro. Error propagation in machine vision. MVA, 7:93–114, 1994.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2000

Authors and Affiliations

  • Yvan G. Leclerc
    • 1
  • Q.-Tuan Luong
    • 1
  • P. Fua
    • 2
  1. 1.Artificial Intelligence CenterSRI InternationalMenlo Park
  2. 2.LIG, EPFLLausanneSwitzerland

Personalised recommendations