Automatic singularity test for motion analysis by an information criterion

  • Kenichi Kanatani
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1064)


The structure-from-motion algorithm from two views fails if the object is a planar surface or the camera motion is a pure rotation. This paper presents a new scheme for automatically detecting these anomalies without using any knowledge about the noise in the images. This judgment does not involve any empirically adjustable thresholds, either. The basic principle of our scheme is to choose a model that has “higher predicting capability” measured by the geometric information criterion (geometric AIC).


Feature Point Planar Surface Maximum Likelihood Estimator Motion Parameter Generalize Inverse 
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.
    Akaike, H.: A new look at the statistical model identification. IEEE Trans. Automation Control 19-6 (1974) 776–723Google Scholar
  2. 2.
    Faugeras, O.: Three-Dimensional Computer Vision: A Geometric Viewpoint. MIT Press, Cambridge, MA, 1993Google Scholar
  3. 3.
    Faugeras, O. D. and S. Maybank, S.: Motion from point matches: Multiplicity of solutions. Int. J. Comput. Vision 4-3 (1990) 225–246Google Scholar
  4. 4.
    Kanatani, K.: Geometric Computation for Machine Vision. Oxford University Press, Oxford, 1993Google Scholar
  5. 5.
    Kanatani, K.: Renormalization for motion analysis: Statistically optimal algorithm. IEICE Trans. Inf. & Sys. E77-D-11 (1994) 1233–1239Google Scholar
  6. 6.
    Kanatani, K.: Statistical Optimization for Geometric Computation: Theory and Practice. Elsevier Science, Amsterdam, 1996Google Scholar
  7. 7.
    Kanatani, K. and Takeda, S.: 3-D motion analysis of a planar surface by renormalization. IEICE Trans. Inf. & Syst. E78-D-8 (1995) 1074–1079Google Scholar
  8. 8.
    Longuet-Higgins, L. C.: The reconstruction of a plane surface from two perspective projections. Proc. Roy. Soc. Lond. B227 (1986) 399–410Google Scholar
  9. 9.
    Maybank, S.: Theory of Reconstruction from Image Motion. Springer, Berlin, 1993Google Scholar
  10. 10.
    Weng, J., Ahuja, N. and Huang, T. S.: Motion and structure from point correspondences with error estimation: Planar surfaces. IEEE Trans. Sig. Proc. 39-12 (1991) 2691–2717Google Scholar
  11. 11.
    Weng, J., Ahuja, N. and Huang, T. S.: Optimal motion and structure estimation. IEEE Trans. Patt. Anal. Mach. Intell. 15-9 (1993) 864–884Google Scholar
  12. 12.
    Weng, J., Huang, T. S. and Ahuja, N.: Motion and Structure from Image Sequences. Springer, Berlin, 1993Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1996

Authors and Affiliations

  • Kenichi Kanatani
    • 1
  1. 1.Department of Computer ScienceGunma UniversityGunmaJapan

Personalised recommendations