Shape Modeling by Optimising Description Length Using Gradients and Parameterisation Invariance

Conference paper
Part of the Springer Proceedings in Mathematics book series (PROM, volume 6)

Abstract

In Statistical Shape Modeling, a dense correspondence between the shapes in the training set must be established. Traditionally this has been done by hand, a process that commonly requires a lot of work and is difficult, especially in 3D. In recent years there has been a lot of work on automatic construction of Shape Models. In recent papers (Davies et al., Medical Image Computing and Computer-Assisted Intervention MICCAI’2001, pp. 57–65, 2001; Davies et al., IEEE Trans. Med. Imaging. 21(5):525–537 2002; Kotcheff and Taylor, Med. Image Anal. 2:303–314 1998) Minimum Description Length, (MDL), is used to locate a dense correspondence between shapes. In this paper the gradient of the description length is derived. Using the gradient, MDL is optimised using steepest descent. The optimisation is therefore faster and experiments show that the resulting models are better. To characterise shape properties that are invariant to similarity transformations, it is first necessary to normalise with respect to the similarity transformations. This is normally done using Procrustes analysis. In this paper we propose to align shapes using the MDL criterion. The MDL based algorithm is compared to Procrustes on a number of data sets. It is concluded that there is improvement in generalisation when using MDL to align the shapes. In this paper novel theory to prevent the commonly occurring problem of clustering under correspondence optimisation is also presented. The problem is solved by calculating the covariance matrix of the shapes using a scalar product that is invariant to mutual reparameterisations. An algorithm for implementing the ideas is proposed and compared to Thodberg’s state of the art algorithm for automatic shape modeling. The suggested algorithm is more stable and the resulting models are of higher quality according to the generalisation measure and according to visual inspection of the specificity.

Keywords

Scalar Product Shape Mode Parameterisation Function Steep Descent Shape Model 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Notes

Acknowledgements

We thank H. Thodberg (IMM, DTU) for the silhouettes and B. Kimia et.al. for the images of the sharks, birds, flight birds, rats and forks.

This work has been financed by the SSF sponsored project ‘Vision in Cognitive Systems’ (VISCOS) and by UMAS and the Swedish Knowledge Foundation through the Industrial PhD program in Medical Bioinformatics at the Centre for Medical Innovations (CMI) at the Karolinska Institute.

References

  1. 1.
    Andrew, A., Chu, E., Lancaster, P.: Derivatives of eigenvalues and eigenvectors of matrix functions. SIAM93 J. Matrix Anal. Appl. 14, 903–926 (1993)Google Scholar
  2. 2.
    Baumberg, A., Hogg, D.: Learning flexible models from image sequences. In: Proceedings of the European Conference on Computer Vision, ECCV’94, pp. 299–308. (1994)Google Scholar
  3. 3.
    Belongie, S., Malik, J., Puzicha, J.: Shape matching and object recognition using shape contexts. IEEE Trans. Pattern Anal. Mach. Intell. 24(24), 509–522 (2002)CrossRefGoogle Scholar
  4. 4.
    Benayoun, A., Ayache, N., Cohen, I.: Adaptive meshes and nonrigid motion computation. In: Proceedings of the International Conference on Pattern Recognition, ICPR’94, pp. 730–732. (1994)Google Scholar
  5. 5.
    Bookstein, F.L.: Size and shape spaces for landmark data in two dimensions. Statist. Sci. 1(2), 181–242 (1986)CrossRefMATHGoogle Scholar
  6. 6.
    Bookstein, F.L.: Landmark methods for forms without landmarks: Morphometrics of group differences in outline shape. Med. Image Anal. 3, 225–243 (1999)Google Scholar
  7. 7.
    Chui, H., Rangarajan, A.: A feature registration framework using mixture models. In: IEEE Workshop on Mathematical Methods in Biomedical Image Analysis (MMBIA), pp. 190–197. (2000)Google Scholar
  8. 8.
    Chui, H., Rangarajan, A.: A new algorithm for non-rigid point matching. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), vol. II, pp. 44–51. (2000)Google Scholar
  9. 9.
    Davies, R.H.: Learning shape: Optimal models for analysing natural variability. Ph.D. thesis, Division of Imaging Science and Biological Engineering, University of Manchester, Manchester, England (2002)Google Scholar
  10. 10.
    Davies, R.H., Cootes, T.F., Taylor, C.J.: A minimum description length approach to statistical shape modeling. In: Information Processing in Medical Imaging, pp. 50–63. (2001)Google Scholar
  11. 11.
    Davies, R.H., Cootes, T.F., Waterton, J.C., Taylor, C.J.: An efficient method for constructing optimal statistical shape models. In: Medical Image Computing and Computer-Assisted Intervention MICCAI’2001, pp. 57–65. (2001)Google Scholar
  12. 12.
    Davies, R.H., Twining, C.J., Cootes, T.F., Allen, P.D., Taylor, C.J.: Shape discrimination in the hippocampus using an mdl model. In: Information Processing in Medical Imaging, pp. 684–695. (2003)Google Scholar
  13. 13.
    Davies, R.H., Twining, C.J., Cootes, T.F., Waterton, J.C., Taylor, C.J.: A minimum description length approach to statistical shape modeling. IEEE Trans. Med. Imaging 21(5), 525–537 (2002)CrossRefGoogle Scholar
  14. 14.
    Dryden, I.L., Mardia, K.V.: Statistical Shape Analysis. Wiley, New York (1999)Google Scholar
  15. 15.
    Ericsson, A.: Automatic shape modelling and applications in medical imaging. Licentiate thesis, Department of Mathematics, Lund Institute of Technology, Centre for Mathematical Sciences, Lund, Sweden (2003)Google Scholar
  16. 16.
    Ericsson, A., Karlsson, J.: Aligning shapes by minimising the description length. In: Proceedings of the Scandinavian Conference on Image Analysis, SCIA’05, vol. 3540/2005, pp. 709–718. Joensuu, Finland (2005)Google Scholar
  17. 17.
    Ericsson, A., Karlsson, J.: Benchmarking of algorithms for automatic correspondence localisation. In: Proceedings of the British Machine Vision Conference, BMVC’06. Edinburgh, Great Britain, Sept 2011Google Scholar
  18. 18.
    Ericsson, A., Åström, K.: An affine invariant deformable shape representation for general curves. In: Proceedings of the International Conference on Computer Vision, ICCV’03, Nice, France (2003)Google Scholar
  19. 19.
    Ericsson, A., Åström, K.: Minimizing the description length using steepest descent. In: Proceedings of the British Machine Vision Conference, Norwich, United Kingdom (2003)Google Scholar
  20. 20.
    Forsyth, D.A., Ponce, J.: Computer Vision. Prentice Hall, New Jersey (2003)Google Scholar
  21. 21.
    Gower, J.C.: Generalized procrustes analysis. Psychometrica 40, 33–50 (1975)MATHMathSciNetGoogle Scholar
  22. 22.
    Hill, A., Taylor, C.J.: Automatic landmark generation for point distribution models. In: Proceedings of the British Machine Vision Conference, pp. 429–438 (1994)Google Scholar
  23. 23.
    Hill, A., Taylor, C.J.: A framework for automatic landmark indentification using a new method of nonrigid correspondence. IEEE Trans. Pattern Anal. Mach. Intell. 22, 241–251 (2000)CrossRefGoogle Scholar
  24. 24.
    Kambhamettu, C., Goldgof, D.B.: Points correspondences recovery in non-rigid motion. In: Proceedings of the Conference on Computer Vision and Pattern Recognition, CVPR’92, pp. 222–237 (1992)Google Scholar
  25. 25.
    Karlsson, J., Ericsson, A.: A geodesic ground truth correspondence measure for benchmarking. In: Proceedings of the International Conference on Pattern Recognition, ICPR’06. Hong Kong, China, Aug 2011Google Scholar
  26. 26.
    Karlsson, J., Ericsson, A., Åström, K.: Parameterisation invariant statistical shape models. In: Proceedings of the International Conference on Pattern Recognition, Cambridge, UK (2004)Google Scholar
  27. 27.
    Kelemen, A., Szekely, G., Gerig, G.: Elastic model-based segmentation of 3d neuroradiological data sets. IEEE Trans. Med. Imaging 18(10), 828–839 (1999)CrossRefGoogle Scholar
  28. 28.
    Kendall, D.G., Barden, D., Carne, T.K., Le, H.: Shape and Shape Theory. Wiley, Chichester (1999)CrossRefMATHGoogle Scholar
  29. 29.
    Kotcheff, A.C.W., Taylor, C.J.: Automatic construction of eigenshape models by direct optimization. Med. Image Anal. 2, 303–314 (1998)CrossRefGoogle Scholar
  30. 30.
    Ljung, L.: System Identification. Prentice Hall, New Jersey (1998)Google Scholar
  31. 31.
    Papadopoulo, T., Lourakis, M.: Estimating the jacobian of the singular value decomposition. In: Proceedings of the European Conference on Computer Vision, ECCV’00, pp. 555–559. (2000)Google Scholar
  32. 32.
    Rissanen, J.: Modeling by shortest data description. Automatica 14, 465–471 (1978)MATHGoogle Scholar
  33. 33.
    Rueckert, D., Frangi, F., Schnabel, J.A.: Automatic construction of 3d-statistical deformation models using nonrigid registration. In: Medical Image Computing and Computer-Assisted Intervention MICCAI’2001, pp. 77–84. (2001)Google Scholar
  34. 34.
    Sebastian, T., Klein, P., Kimia, B.: Constructing 2d curve atlases. In: IEEE Workshop on Mathematical Methods in Biomedical Image Analysis, pp. 70–77. (2000)Google Scholar
  35. 35.
    Shannon, C.E.: A mathematical theory of communication. Bell Syst. Techn. J. 27, 379–423; 623–656 (1948)MathSciNetGoogle Scholar
  36. 36.
    Small, C.G.: The Statistical Theory of Shape. Springer, New York, (1996)CrossRefMATHGoogle Scholar
  37. 37.
    Tagare, H.D.: Shape-based nonrigid correspondence with application to heart motion analysis. IEEE Trans. Med. Imaging 18, 570–579 (1999)CrossRefGoogle Scholar
  38. 38.
    Thodberg, H.H.: Minimum description length shape and appearance models. In: Image Processing Medical Imaging, IPMI’03, pp. 51–62. (2003)Google Scholar
  39. 39.
    Thodberg, H.H.: Minimum description length shape and appearance models. Technical report: IMM Technical Report 2003-01, Technical University of Denmark, Lyngby (2003)Google Scholar
  40. 40.
    Wang, Y., Peterson, B.S., Staib, L.H.: Shape-based 3d surface correspondence using geodesics and local geometry. In: Proceedings of the Conference on Computer Vision and Pattern Recognition, CVPR’00, pp. 644–651. (2000)Google Scholar
  41. 41.
    Zheng, Y., Doermann, D.: Robust point matching for non-rigid shapes: A relaxation labeling based approach. Technical report: Lamp-tr-117, University of Maryland, College Park (2004)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Johan Karlsson
    • 1
  • Anders Ericsson
    • 2
  • Kalle Åström
    • 2
  1. 1.Fraunhofer-Chalmers Research Centre for Industrial MathematicsGöteborgSweden
  2. 2.Center for Mathematical SciencesLund UniversityLundSweden

Personalised recommendations