Pose Priors for Simultaneously Solving Alignment and Correspondence

  • Francesc Moreno-Noguer
  • Vincent Lepetit
  • Pascal Fua
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5303)


Estimating a camera pose given a set of 3D-object and 2D-image feature points is a well understood problem when correspondences are given. However, when such correspondences cannot be established a priori, one must simultaneously compute them along with the pose. Most current approaches to solving this problem are too computationally intensive to be practical. An interesting exception is the SoftPosit algorithm, that looks for the solution as the minimum of a suitable objective function. It is arguably one of the best algorithms but its iterative nature means it can fail in the presence of clutter, occlusions, or repetitive patterns. In this paper, we propose an approach that overcomes this limitation by taking advantage of the fact that, in practice, some prior on the camera pose is often available. We model it as a Gaussian Mixture Model that we progressively refine by hypothesizing new correspondences. This rapidly reduces the number of potential matches for each 3D point and lets us explore the pose space more thoroughly than SoftPosit at a similar computational cost. We will demonstrate the superior performance of our approach on both synthetic and real data.


Object Recognition Gaussian Mixture Model Gaussian Component Repetitive Pattern Uncertainty Region 
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.
    David, P., DeMenthon, D., Duraiswami, R., Samet, H.: Softposit: Simultaneous pose and correspondence determination. IJCV 59, 259–284 (2004)CrossRefzbMATHGoogle Scholar
  2. 2.
    Fischler, M.A., Bolles, R.C.: Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography. Commun. ACM 24, 381–395 (1981)MathSciNetCrossRefGoogle Scholar
  3. 3.
    Grimson, W.E.L., Meyer-Arendt, J.R.: Object recognition by computer: the role of geometric constraints, vol. 31. MIT Press, Cambridge (1992)Google Scholar
  4. 4.
    Grimson, W.E.L.: The combinatorics of heuristic search termination for object recognition in cluttered environments. IEEE PAMI 13, 920–935 (1991)CrossRefGoogle Scholar
  5. 5.
    Breuel, T.: Fast recognition using adaptive subdivisions of transformation space. In: Proc. IEEE CVPR, pp. 445–451 (1992)Google Scholar
  6. 6.
    Cass, T.A.: Polynomial-time object recognition in the presence of clutter, occlusion, and uncertainty. In: Proc. ECCV, pp. 834–842 (1992)Google Scholar
  7. 7.
    Olson, C.F.: Efficient pose clustering using a randomized algorithm. IJCV 23, 131–147 (1997)CrossRefGoogle Scholar
  8. 8.
    Stockman, G.: Object recognition and localization via pose clustering. Comput. Vision Graph. Image Process 40, 361–387 (1987)CrossRefGoogle Scholar
  9. 9.
    Ayache, N., Faugeras, O.D.: Hyper: a new approach for the recognition and positioning to two-dimensional objects. IEEE PAMI 8, 44–54 (1986)CrossRefGoogle Scholar
  10. 10.
    Grimson, W.E.L., Lozano-Pérez, T.: Localizing overlapping parts by searching the interpretation tree. IEEE PAMI 9, 469–482 (1987)CrossRefGoogle Scholar
  11. 11.
    Lamdan, Y., Wolfson, H.J.: Geometric hashing: A general and efficient model-based recognition scheme. In: Proc. IEEE ICCV, pp. 238–249 (1988)Google Scholar
  12. 12.
    Burns, J.B., Weiss, R.S., Riseman, E.M.: View variation of point-set and line-segment features. IEEE PAMI 15, 51–68 (1993)CrossRefGoogle Scholar
  13. 13.
    Beis, J.S., Lowe, D.G.: Indexing without invariants in 3D object recognition. IEEE PAMI 21, 1000–1015 (1999)CrossRefGoogle Scholar
  14. 14.
    Jurie, F.: Solution of the simultaneous pose and correspondence problem using gaussian error model. CVIU 73, 357–373 (1999)zbMATHGoogle Scholar
  15. 15.
    Wunsch, P., Hirzinger, G.: Registration of cad-models to images by iterative inverse perspective matching. In: Proc. ICPR., vol. 78 (1996)Google Scholar
  16. 16.
    Beveridge, J.R., Riseman, E.M.: Optimal geometric model matching under full 3D perspective. CVIU 61, 351–364 (1995)Google Scholar
  17. 17.
    Arya, S., Mount, D., Netanyahu, N., Silverman, R., Wu, A.: An optimal algorithm for approximate NN searching fixed dimensions. J. ACM 45, 891–923 (1998)MathSciNetCrossRefzbMATHGoogle Scholar
  18. 18.
    David, P., DeMenthon, D.: Object recognition in high clutter images using line features. In: Proc. IEEE ICCV, pp. 1581–1588 (2005)Google Scholar
  19. 19.
    Lowe, D.G.: Distinctive image features from scale-invariant keypoints. IJCV 15, 91–110 (2004)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Francesc Moreno-Noguer
    • 1
  • Vincent Lepetit
    • 1
  • Pascal Fua
    • 1
  1. 1.Computer Vision Laboratory École Polytechnique Fédérale de Lausanne (EPFL)LausanneSwitzerland

Personalised recommendations