Automatic Mutual Nonrigid Registration of Dense Surfaces by Graphical Model Based Inference

  • Xiao Dong
  • Guoyan Zheng
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5342)


This paper addresses the problem of fully automatic matching two triangulated surface meshes. In this paper, a similarity measurement is constructed to measure the consistency of the constraints among the correspondent landmarks, which is rigid transformation immune and robust to nonrigid deformations. The matching problem is then solved by directly finding correspondence between the landmarks of the two surfaces by graphical model based Bayesian inference. In order to reduce the computational complexity and to accelerate the convergence, a hierarchical graphical model is constructed which enables mutual registration and information exchange between the two surfaces during registration. Experiments on randomly generated instances from a PCA based statistical model of proximal femurs verified the proposed approach.


Graphical Model Belief Propagation Shape Descriptor Coarse Level Rigid Transformation 
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.
    Cootes, T., Taylor, C.: Statistical models of appearance for computer vision. Technical report, University of Manchester, U.K (2004)Google Scholar
  2. 2.
    Xu, C., Yezzi, A., Prince, J.: A summary of geometric level-set analogues for a general class of parametric active contour and surface models (2001)Google Scholar
  3. 3.
    Lee, S.M., Abbott, A.L., Clark, N.A., Araman, P.A.: A shape representation for planar curves by shape signature harmonic embedding. In: CVPR 2006, pp. 1940–1947 (2006)Google Scholar
  4. 4.
    Roy, A.S., Gopinath, A., Rangarajan, A.: Deformable density matching for 3d non-rigid registration of shapes. In: MICCAI 2007, pp. 942–949 (2007)Google Scholar
  5. 5.
    Jiang, Y.F., Xie, J., Sun, D.Q., Tsui, H.: Shape registration by simultaneously optimizing representation and transformation. In: MICCAI 2007, pp. 809–817 (2007)Google Scholar
  6. 6.
    Belongie, S., Malik, J., Puzicha, J.: Shape matching and object recognition using shape contexts. IEEE Transactions on Pattern Analyis and Machine Intelligence 24, 509–522 (2002)CrossRefGoogle Scholar
  7. 7.
    Jain, V., Zhang, H.: Robust 3d shape correspondence in the spectral domain. In: International Conference on Shape Modeling and Applications, SMI (2006)Google Scholar
  8. 8.
    Coughlan, J., Ferreira, S.: Finding deformable shapes using loopy belief propagation. In: ECCV 2003, pp. 453–468 (2002)Google Scholar
  9. 9.
    Caetano, T.S., Caeli, T., Barone, D.A.C.: An optimal probabilistic graphical model for point set matching. Technical Report Technical Report TR 04-03, University of Alberta, Edmonton, Alberta Canada (2004)Google Scholar
  10. 10.
    Rangarajan, A., Coughlan, J., Yuille, A.L.: A bayesian network framework for relational shape matching. In: ICCV 2003, pp. 671–678 (2003)Google Scholar
  11. 11.
    Zhang, L., Seitz, S.M.: Parameter estimation for mrf stereo. In: CVPR 2005, pp. 288–295 (2005)Google Scholar
  12. 12.
    Sun, J., Zheng, N.N., Shum, H.Y.: Stereo matching using belief propagation. IEEE Transactions on Pattern Analysis and Machine Interlligence 25, 1–14 (2003)CrossRefzbMATHGoogle Scholar
  13. 13.
    Xiao, P.D., Barnes, N., Caetano, T., Lieby, P.: An MRF and Gaussian curvature based shape representation for shape matching. In: CVPR 2007, pp. 17–22 (2007)Google Scholar
  14. 14.
    Gibbs, A.L.: Bounding the convergence time of the gibbs sampler in bayesian image restroation. Biometrika 87(4), 749–766 (2000)MathSciNetCrossRefzbMATHGoogle Scholar
  15. 15.
    McEliece, R.J., MacKay, D.J.C., Cheng, J.F.: Turbo decoding as an instance of pearl’s ”beliefpropagation” algorithm. IEEE Journal on Selected Areas in Communications 16, 140–152 (1998)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Xiao Dong
    • 1
  • Guoyan Zheng
    • 1
  1. 1.MEM Research CenterUniversity of BernSwitzerland

Personalised recommendations