Advertisement

IMPSAC: Synthesis of Importance Sampling and Random Sample Consensus

  • P. H. S. Torr
  • C. Davidson
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1843)

Abstract

This paper proposes a new method for effecting feature correspondence between images. The method operates from coarse to fine and is superior to previous methods in that it can solve the wide baseline stereo problem, even when the image has been deformed or rotated. At the coarsest level a RANSAC-style estimator is used to estimate the two view image constraint R which is then used to guide matching. The two view relation is an augmented fundamental matrix, being a fundamental matrix plus a homography consistent with that fundamental matrix. This is akin to the plane plus parallax representation with the homography being used to help guide matching and to mitigate the effects of image deformation. In order to propagate the information from coarse to fine images, the distribution of the parameters Θ of R is encoded using a set of particles and an importance sampling function. It is not known in general how to choose the importance sampling function, but a new method “IMPSAC” is presented that automatically generates such a function. It is shown that the method is superior to previous single resolution RANSAC-style feature matchers.

Keywords:

Structure from motion Stereoscopic vision 

References

  1. 1.
    Ayache N. Artificial vision for mobile robots. MIT Press, Cambridge, 1991.Google Scholar
  2. 2.
    Beardsley P., Torr P., and Zisserman A. 3D model acquisition from extended image sequences. In Proc. European Conference on Computer Vision, LNCS 1064/1065, pages 683–695. Springer-Verlag, 1996.Google Scholar
  3. 3.
    Beardsley P., Zisserman A., and Murray D. Navigation using affine structure and motion. In Proc. European Conference on Computer Vision, LNCS800/801, pages 85–96. Springer-Verlag, 1994.Google Scholar
  4. 4.
    J. R. Bergen, P. Anandan, K. Hanna, and R. Hingorani. Hierarchical model-based motion estimation. In Proc. 2nd European Conference on ComputerVisionLNCS588, Santa Margherita Ligure, pages 237–252, 1992.Google Scholar
  5. 5.
    T. Cham and R. Cipolla. A statistical framework for long range matching in uncalibrated image mosaicing. In Conference on Computer Vision and Pattern Recognition, pages 442–447, 1998.Google Scholar
  6. 6.
    A. P. Dempster, N. M. Laird, and D. B. Rubin. Maximum likelihood from incomplete data via the em algorithm. J. R. Statist. Soc., 39 B:1–38, 1977.MathSciNetGoogle Scholar
  7. 7.
    O.D. Faugeras. What can be seen in three dimensions with an uncalibrated stereo rig? In G. Sandini, editor, Proc. 2nd European Conference on Computer Vision, LNCS588, Santa Margherita Ligure, pages 563–578. Springer-Verlag, 1992.Google Scholar
  8. 8.
    M. Fischler and R. Bolles. Random sample consensus: a paradigm for model fitting with application to image analysis and automated cartography. Commun. Assoc. Comp. Mach., vol. 24:381–95, 1981.MathSciNetGoogle Scholar
  9. 9.
    A. Gelman, J. Carlin, H. Stern, and D. Rubin. Bayesian Data Analysis. Chapman and Hall, 1995.Google Scholar
  10. 10.
    C. Harris and M. Stephens. A combined corner and edge detector. In Proc. Alvey Conf., pages 189–192, 1987.Google Scholar
  11. 11.
    Harris C. Determination of ego-motion from matched points. In Third Alvey Vision Conference, pages 189–192, 1987.Google Scholar
  12. 12.
    R. I. Hartley. Estimation of relative camera positions for uncalibrated cameras. In Proc. 2nd European Conference on Computer Vision, LNCS 588, Santa Margherita Ligure, pages 579–587. Springer-Verlag, 1992.Google Scholar
  13. 13.
    Irani M. and Anandan P. Parallax geometry of pairs of points for 3d scene analysis. In Buxton B. and Cipolla R., editors, Proc. 4th European Conference on Computer Vision, LNCS 1064, Cambridge, pages 17–30. Springer, 1996.Google Scholar
  14. 14.
    M. Isard and A. Blake. Condensation-conditional density propagation for visual tracking. International Journal of Computer Vision, 28(1):5–28, 1998.CrossRefGoogle Scholar
  15. 15.
    E. T. Jaynes. Probability theory as extended logic. Not yet published a postscript version of this excellent book is available at ftp://bayes.wustl.edu/pub/Jaynes/, 1999.
  16. 16.
    G.I. McLachlan and K. Basford. Mixture models: inference and applications to clustering. Marcel Dekker. New York, 1988.zbMATHGoogle Scholar
  17. 17.
    McLauchlan P. and Murray D. A unifying framework for structure from motion recovery from image sequences. In Proc. International Conference on Computer Vision, pages 314–320, 1995.Google Scholar
  18. 18.
    J. Mundy and A. Zisserman. Geometric Invariance in Computer Vision. MIT press, 1992.Google Scholar
  19. 19.
    P. Pritchett and A. Zisserman. Wide baseline stereo matching. In Proc. 6th International Conference on Computer Vision, Bombay, pages 754–760, January 1998.Google Scholar
  20. 20.
    L. S. Shapiro. Affine Analysis of Image Sequences. PhD thesis, Oxford University, 1993.Google Scholar
  21. 21.
    J. Sullivan, A. Blake, M. Isard, and J. MacCormick. Bayesian correlation. In Seventh International Conference on Computer Vision, volume 2, pages 1068–1075, 1999.CrossRefGoogle Scholar
  22. 22.
    C. Tomasi and T. Kanade. Shape and motion from image streams under orthography: A factorisation approach. International Journal of Computer Vision, 9(2): 137–154, 1992.CrossRefGoogle Scholar
  23. 23.
    P. H. S. Torr. Outlier Detection and Motion Segmentation. PhD thesis, Dept. of Engineering Science, University of Oxford, 1995.Google Scholar
  24. 24.
    P. H. S. Torr and D. W. Murray. Outlier detection and motion segmentation. In P. S. Schenker, editor, Sensor Fusion VI, pages 432–443. SPIE volume 2059, 1993. Boston.Google Scholar
  25. 25.
    P. H. S. Torr and D. W. Murray. The development and comparison of robust methods for estimating the fundamental matrix. Int Journal of Computer Vision, 24(3):271–300, 1997.CrossRefGoogle Scholar
  26. 26.
    P. H. S. Torr and A. Zisserman. Robust computation and parametrization of multiple view relations. In U Desai, editor, ICCV6, pages 727–732. Narosa Publishing House, 1998.Google Scholar
  27. 27.
    P.H.S. Torr, A. Fitzgibbon, and A. Zisserman. The problem of degeneracy in structure and motion recovery from uncalibrated image sequences. IJCV, 32(1):27–45, 1999. Marr Prize Paper ICCV 1999.CrossRefGoogle Scholar
  28. 28.
    Zeller, C. Projective, Affine andEuclidean Calibration in Compute Vision and the Application of Three Dimensional Perception. PhD thesis, Robot Vis Group, INRIA Sophia-Antipolis, 1996.Google Scholar
  29. 29.
    Z. Zhang, R. Deriche, O. Faugeras, and Q. T. Luong. A robust technique for matching two uncalibrated images through the recovery of the unknown epipolar geometry. AI Journal, vol. 78:87–119, 1994.Google Scholar
  30. 30.
    Z. Zhang and O. Faugeras. 3D Dynamic Scene Analysis. Springer-Verlag, 1992.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2000

Authors and Affiliations

  • P. H. S. Torr
    • 1
  • C. Davidson
    • 1
  1. 1.Microsoft Research LtdCambridgeUK

Personalised recommendations