Advertisement

Enforcing a Shape Correspondence between Two Views of a 3D Non-rigid Object

  • M. Benjamin Dias
  • Bernard F. Buxton
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2905)

Abstract

We have developed an algorithm capable of enforcing a shape correspondence between two views of the same object in different shape-states. This algorithm, together with several other significant updates, has helped improve the performance of the Integrated Shape and Pose Model (ISPM) described in [1] by a factor of 10. The ISPM utilizes two flexible basis views to integrate the linear combination of views technique with a coupled-view Flexible Shape Model (FSM) [2]. As a proof-of-principle we have evaluated the performance of the improved ISPM in comparison to that of its predecessor [1] and of the conventional FSM [3], via two different databases. The results show that, unlike the FSM, the current ISPM is view-invariant and that, on average, it out-performs the FSM. It also out-performs the initial ISPM described in [1].

References

  1. 1.
    Dias, M.B., Buxton, B.F.: Integrated shape and pose modelling. In: Proc. BMVC, vol. 2, pp. 827–836 (2002)Google Scholar
  2. 2.
    Cootes, T.F., Wheeler, G.V., Walker, K.N., Taylor, C.J.: Coupled-view active appearance models. In: Proc. BMVC, vol. 1, pp. 52–61 (2000)Google Scholar
  3. 3.
    Cootes, T.F., Taylor, C.J., Cooper, D.H., Graham, J.: A trainable method of parametric shape description. Image and Vision Computing 10, 289–294 (1992)CrossRefGoogle Scholar
  4. 4.
    Bretzner, L., Lindeberg, T.: Use your hands as a 3-d mouse, or, relative orientation from extended sequences of sparse point and line correspondences using the affine trifocal tensor. In: Burkhardt, H.-J., Neumann, B. (eds.) ECCV 1998. LNCS, vol. 1406, pp. 141–157. Springer, Heidelberg (1998)Google Scholar
  5. 5.
    Tarr, M.J., Williams, P., Hayward, W.G., Gauthier, I.: Three-dimensional object recognition is viewpoint dependent. Nature Neuroscience 1, 275–277 (1998)CrossRefGoogle Scholar
  6. 6.
    Ullman, S., Basri, R.: Recognition by linear combinations of models. IEEE Transactions on PAMI 13, 992–1006 (1991)Google Scholar
  7. 7.
    Pollard, S., Pilu, M., Hayes, S., Lorusso, A.: View synthesis by trinocular edge matching and transfer. In: Proc. BMVC, vol. 2, pp. 770–779 (1998)Google Scholar
  8. 8.
    Koufakis, I., Buxton, B.F.: Very low bit-rate face video compression using linear combination of 2d face views and principal component analysis. Image and Vision Computing 17, 1031–1051 (1998)CrossRefGoogle Scholar
  9. 9.
    Thórhallsson, T., Murray, D.W.: The tensors of three affine views. In: Proc. International Conference on Computer Vision and Pattern Recognition, vol. 1, pp. 450–456 (1999)Google Scholar
  10. 10.
    Fischler, M.A., Bolles, R.C.: Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography. Communications of the ACM 24, 381–395 (1981)MathSciNetCrossRefGoogle Scholar
  11. 11.
    Hartley, R., Zisserman, A.: Multiple View Geometry In Computer Vision. Cambridge University Press, Cambridge (2000) ISBN 0-521-62304-9zbMATHGoogle Scholar
  12. 12.
    Loizides, A., Slater, M., Langdon, W.B.: Measuring facial emotional expressions using genetic programming. In: Soft Computing and Industry Recent Applications, pp. 545–554 (2001)Google Scholar
  13. 13.
    Dias, M.B., Buxton, B.F.: Advances in integrated shape and pose modelling. Technical Report RN/03/10, University College London (UCL), Department of Computer Science (2003)Google Scholar
  14. 14.
    Kohavi, R.: A study of cross-validation and bootstrap for accuracy estimation and model selection. In: Proc. IJCAI (1995)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • M. Benjamin Dias
    • 1
  • Bernard F. Buxton
    • 1
  1. 1.Department of Computer ScienceUniversity College LondonLondonUK

Personalised recommendations