Handling Missing Data in the Computation of 3D Affine Transformations

  • Hanna Martinsson
  • Adrien Bartoli
  • François Gaspard
  • Jean-Marc Lavest
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3757)


The reconstruction of rigid scenes from multiple images is a central topic in computer vision. Approaches merging partial 3D models in a hierarchical manner have proven the most effective to deal with large image sequences. One of the key building blocks of these hierarchical approaches is the alignment of two partial 3D models, which requires to express them in the same 3D coordinate frame by computing a 3D transformation. This problem has been well-studied for the cases of 3D models obtained with calibrated or uncalibrated pinhole cameras.

We tackle the problem of aligning 3D models – sets of 3D points – obtained using uncalibrated affine cameras. This requires to estimate 3D affine transformations between the models. We propose a factorization-based algorithm estimating simultaneously the aligning transformations and corrected points, exactly matching the estimated transformations, such that the reprojection error over all cameras is minimized. In the case of incomplete image data our algorithm uses an Expectation Maximization (EM) based scheme that alternates prediction of the missing data and estimation of the affine transformation.

We experimentally compare our algorithm to other methods using simulated and real data.


Image Point Point Correspondence Factorization Algorithm Reprojection Error Corrected Point 
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.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Fitzgibbon, A., Zisserman, A.: Automatic camera recovery for closed or open image sequences. In: ECCV, pp. 311–326 (1998)Google Scholar
  2. 2.
    Tomasi, C., Kanade, T.: Shape and motion from image streams under orthography: A factorization method. IJCV 9, 137–154 (1992)CrossRefGoogle Scholar
  3. 3.
    Beardsley, P., Zisserman, A., Murray, D.: Sequential updating of projective and affine structure from motion. IJCV 23, 235–259 (1997)CrossRefGoogle Scholar
  4. 4.
    Csurka, G., Demirdjian, D., Horaud, R.: Finding the collineation between two projective reconstructions. Comp. Vision and Image Underst. 75, 260–268 (1999)CrossRefGoogle Scholar
  5. 5.
    Walker, M., Shao, L., Volz, R.: Estimating 3D location parameters using dual number quaternions. Computer Vision, Graphics and Image Processing: Image Understanding 54, 358–367 (1991)zbMATHGoogle Scholar
  6. 6.
    Mundy, J., Zisserman, A. (eds.): Geometric Invariance in Computer Vision. MIT Press, Cambridge (1992)Google Scholar
  7. 7.
    Bartoli, A., Martinsson, H., Gaspard, F., Lavest, J.M.: On aligning sets of points reconstructed from uncalibrated affine cameras. In: SCIA, pp. 531–540 (2005)Google Scholar
  8. 8.
    Reid, I., Murray, D.: Active tracking of foveated feature clusters using affine structure. IJCV 18, 41–60 (1996)CrossRefGoogle Scholar
  9. 9.
    Golub, G., van Loan, C.: Matrix Computation. Johns Hopkins University Press, Baltimore (1989)Google Scholar
  10. 10.
    Jacobs, D.: Linear fitting with missing data: Applications to structure-from-motion and to characterizing intensity images. In: CVPR, pp. 206–212 (1997)Google Scholar
  11. 11.
    McLachlan, G., Krishnan, T.: The EM algorithm and extensions. John Wiley & Sons, Inc., Chichester (1997)zbMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2005

Authors and Affiliations

  • Hanna Martinsson
    • 1
  • Adrien Bartoli
    • 2
  • François Gaspard
    • 1
  • Jean-Marc Lavest
    • 2
  1. 1.CEA LIST – LIST/DTSI/SARC/LCEIGif sur YvetteFrance
  2. 2.LASMEA (CNRS / UBP)AubièreFrance

Personalised recommendations