Using Strong Shape Priors for Stereo

  • Yunda Sun
  • Pushmeet Kohli
  • Matthieu Bray
  • Philip H. S. Torr
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4338)


This paper addresses the problem of obtaining an accurate 3D reconstruction from multiple views. Taking inspiration from the recent successes of using strong prior knowledge for image segmentation, we propose a framework for 3D reconstruction which uses such priors to overcome the ambiguity inherent in this problem. Our framework is based on an object-specific Markov Random Field (MRF)[10]. It uses a volumetric scene representation and integrates conventional reconstruction measures such as photo-consistency, surface smoothness and visual hull membership with a strong object-specific prior. Simple parametric models of objects will be used as strong priors in our framework. We will show how parameters of these models can be efficiently estimated by performing inference on the MRF using dynamic graph cuts [7]. This procedure not only gives an accurate object reconstruction, but also provides us with information regarding the pose or state of the object being reconstructed. We will show the results of our method in reconstructing deformable and articulated objects.


Markov Random Field Deformable Model Reconstruction Result Reconstruction Problem Surface Smoothness 
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.
    Bhatia, S., Sigal, L., Isard, M., Black, M.J.: 3d human limb detection using space carving and multi-view eigen models. In: ANM Workshop, vol. I, p. 17 (2004)Google Scholar
  2. 2.
    Boykov, Y., Kolmogorov, V.: An experimental comparison of min-cut/max-flow algorithms for energy minimization in vision. PAMI 26(9), 1124–1137 (2004)Google Scholar
  3. 3.
    Boykov, Y.Y., Jolly, M.P.: Interactive graph cuts for optimal boundary and region segmentation of objects in n-d images. In: ICCV, pp. 105–112 (2001)Google Scholar
  4. 4.
    Bray, M., Kohli, P., Torr, P.H.S.: Posecut: Simulataneous segmentation and 3d pose estimation of humans using dynamic graph cuts. In: ECCV, pp. 642–655 (2006)Google Scholar
  5. 5.
    Meijster, et al.: A general algorithm for computing distance transforms in linear time. In: MMAIS Processing, pp. 331–340 (2000)Google Scholar
  6. 6.
    Press, et al.: Numerical recipes in C. Cambridge Uni. Press, Cambridge (1988)MATHGoogle Scholar
  7. 7.
    Kohli, P., Torr, P.: Efficiently solving dynamic markov random fields using graph cuts. In: ICCV (2005)Google Scholar
  8. 8.
    Kolmogorov, V., Zabih, R.: Multi-camera scene reconstruction via graph cuts. In: Heyden, A., Sparr, G., Nielsen, M., Johansen, P. (eds.) ECCV 2002. LNCS, vol. 2352, p. 82. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  9. 9.
    Kolmogorov, V., Zabih, R.: What energy functions can be minimized via graph cuts? In: Heyden, A., Sparr, G., Nielsen, M., Johansen, P. (eds.) ECCV 2002. LNCS, vol. 2352, p. 65. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  10. 10.
    Kumar, M.P., Torr, P.H.S., Zisserman, A.: Obj cut. In: CVPR, vol. I, pp. 18–25 (2005)Google Scholar
  11. 11.
    Kutulakos, K.N., Seitz, M.: A theory of shape by space carving. IJCV 38(3) (2000)Google Scholar
  12. 12.
    Scharstein, D., Szeliski, R.: A taxonomy and evaluation of dense two-frame stereo correspondence algorithms. IJCV 47(1-3), 7–42 (2002)MATHCrossRefGoogle Scholar
  13. 13.
    Snow, D., Viola, P., Zabih, R.: Exact voxel occupancy with graph cuts. In: CVPR (2000)Google Scholar
  14. 14.
    Szeliski, R.: Rapid octree construction from image sequences. CVGIP 58, 23–32 (1993)CrossRefGoogle Scholar
  15. 15.
    Vogiatzis, G., Torr, P.H.S., Cipolla, R.: Multi-view stereo via volumetric graph-cuts. In: CVPR, vol. II, pp. 391–398 (2005)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Yunda Sun
    • 1
  • Pushmeet Kohli
    • 1
  • Matthieu Bray
    • 1
  • Philip H. S. Torr
    • 1
  1. 1.Department of ComputingOxford Brookes UniversityUK

Personalised recommendations