Skip to main content

Comparison of statistical models performance in case of segmentation using a small amount of training datasets

Abstract

Model-based image segmentation has been extensively used in medical imaging to learn both the shape and appearance of anatomical structures from training datasets. The more training datasets are used, the more accurate is the segmented model, as we account for more information about its variability. However, training datasets of large size with a proper sampling of the population may not always be available. In this paper, we compare the performance of statistical models in the context of lower limb bones segmentation using MR images when only a small number of datasets is available for training. For shape, both PCA-based priors and shape memory strategies are tested. For appearance, methods based on intensity profiles are tested, namely mean intensity profiles, multivariate Gaussian distributions of profiles and multimodal profiles from EM clustering. Segmentation results show that local and simple methods perform the best when a small number of datasets is available for training. Conversely, statistical methods feature the best segmentation results when the number of training datasets is increased.

This is a preview of subscription content, access via your institution.

References

  1. 1.

    Ambroise, C., Dang, M., Govaert, G.: Clustering of spatial data by the em algorithm. Quant. Geol. Geostat. 9, 493–504 (1997)

    Google Scholar 

  2. 2.

    Behiels, G., Maes, F., Vandermeulen, D., Suetens, P.: Evaluation of image features and search strategies for segmentation of bone structures in radiographs using active shape models. Med. Image Anal. 6(1), 47–62 (2002)

    Article  Google Scholar 

  3. 3.

    Chung, F., Delingette, H.: Multimodal prior appearance models based on regional clustering of intensity profiles. In: MICCAI 2009—Proceedings of the 12th International Conference on Medical Image Computing and Computer Assisted Intervention. Lecture Notes in Computer Science, vol. 5762, pp. 1051–1058 (2009)

    Chapter  Google Scholar 

  4. 4.

    Cootes, T., Taylor, C.: Data driven refinement of active shape model search. In: BMVC 1996—Proceedings of the 7th British Machine Vision Conference (1996)

    Google Scholar 

  5. 5.

    Cootes, T., Taylor, C.: (2004) Statistical models of appearance for computer vision, University of Manchester, March 2004

  6. 6.

    Cootes, T.F., Hill, A., Taylor, C.J., Haslam, J.: The use of active shape models for locating structures in medical images. In: IPMI’93—Proceedings of the 13th International Conference on Information Processing in Medical Imaging, pp. 33–47 (1993)

    Chapter  Google Scholar 

  7. 7.

    Cootes, T.F., Edwards, G., Taylor, C.J.: Active appearance models. IEEE Trans. Pattern Anal. Mach. Intell. 23(6), 681–685 (2001)

    Article  Google Scholar 

  8. 8.

    Delingette, H.: General object reconstruction based on simplex meshes. Int. J. Comput. Vis. 32(2), 111–146 (1999)

    Article  Google Scholar 

  9. 9.

    Dice, L.: Measures of the amount of ecologic association between species. Ecology 26(3), 297–302 (1945)

    Article  Google Scholar 

  10. 10.

    Gilles, B., Magnenat-Thalmann, N.: Musculoskeletal MRI segmentation using multi-resolution simplex meshes with medial representations. Med. Image Anal. 14(3), 291–302 (2010)

    Article  Google Scholar 

  11. 11.

    Goodall, C.: Procrustes methods in the statistical analysis of shape. J. R. Stat. Soc., Ser. B, Methodol. 53(2), 285–339 (1991)

    MATH  MathSciNet  Google Scholar 

  12. 12.

    Gower, J.: Generalized Procrustes analysis. Psychometrika 40, 33–51 (1975)

    MATH  Article  MathSciNet  Google Scholar 

  13. 13.

    Heimann, T., Meinzer, H.P.: Statistical shape models for 3d medical image segmentation: a review. Med. Image Anal. 13, 543–563 (2009)

    Article  Google Scholar 

  14. 14.

    Heimann, T., Munzing, S., Meinzer, H., Wolf, I.: A shape-guided deformable model with evolutionary algorithm initialization for 3D soft tissue segmentation. In: Karssemeijer, N., Lelieveldt, B. (eds.) IPMI 2007—Proceedings of the 20th International Conference on Information Processing in Medical Imaging, vol. 4584, pp. 1–12 (2007)

    Google Scholar 

  15. 15.

    Holden, M., Hill, D., Denton, E., Jarosz, J., Cox, T., Hawkes, D.: Voxel similarity measures for 3D serial MR brain image registration. In: IPMI’99—Proceedings of the 16th International Conference on Information Processing in Medical Imaging. Lecture Notes in Computer Science, vol. 1613, pp. 472–477. Springer, Berlin (1999)

    Chapter  Google Scholar 

  16. 16.

    Jolliffe, I.T.: Principal Component Analysis, 2nd edn. Springer, Berlin (2002)

    MATH  Google Scholar 

  17. 17.

    Kim, D.W., Lee, K.H., Lee, D.: On cluster validity index for estimation of the optimal number of fuzzy clusters. Pattern Recognit. 37, 2009–2025 (2004)

    Article  Google Scholar 

  18. 18.

    Schäfer, J., Strimmer, K.: A shrinkage approach to large-scale covariance matrix estimation and implications for functional genomics. Stat. Appl. Genet. Mol. Biol. 4(1), 32 (2005)

    MathSciNet  Google Scholar 

  19. 19.

    Schmid, J., Magnenat-Thalmann, N.: MRI bone segmentation using deformable models and shape priors. In: MICCAI 2008—Proceedings of the 11th International Conference on Medical Image Computing and Computer Assisted Intervention. Lecture Notes in Computer Science, vol. 5241, pp. 119–126 (2008)

    Chapter  Google Scholar 

  20. 20.

    Schmid, J., Sandholm, A., Chung, F., Thalmann, D., Delingette, H., Magnenat-Thalmann, N.: Musculoskeletal Simulation Model Generation from MRI Datasets and Motion Capture Data, pp. 3–19. Springer, Berlin (2009)

    Google Scholar 

  21. 21.

    Schmid, J., Kim, J., Magnenat-Thalmann, N.: Extreme leg motion analysis of professional ballet dancers via MRI segmentation of multiple leg postures. Int. J. Comput. Assist. Radiol. Surg., pp. 1–11. Springer, Berlin (2010)

  22. 22.

    Seim, H., Kainmueller, D., Heller, M., Lamecker, H., Zachow, S., Hege, H.C.: Automatic segmentation of the pelvic bones from CT data based on a statistical shape model. In: Botha, C., Kindlmann, G., Niessen, W., Preim, B. (eds.) Eurographics Workshop on Visual Computing for Biomedicine, Eurographics Association, pp. 93–100 (2008)

    Google Scholar 

  23. 23.

    Styner, M., Brechbühler, C., Székely, G., Gerig, G.: Parametric estimate of intensity inhomogeneities applied to MRI. IEEE Trans. Med. Imaging 19, 153–165 (2000)

    Article  Google Scholar 

  24. 24.

    Tadjudin, S., Landgrebe, D.: Covariance estimation with limited training samples. IEEE Trans. Geosci. Remote Sens. 37(4), 2113–2118 (1999)

    Article  Google Scholar 

Download references

Author information

Affiliations

Authors

Corresponding author

Correspondence to François Chung.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Chung, F., Schmid, J., Magnenat-Thalmann, N. et al. Comparison of statistical models performance in case of segmentation using a small amount of training datasets. Vis Comput 27, 141–151 (2011). https://doi.org/10.1007/s00371-010-0536-9

Download citation

Keywords

  • Model based segmentation
  • Statistical models
  • Principal component analysis
  • Clustering