Advertisement

Learning a Self-organizing Map Model on a Riemannian Manifold

  • D. J. Yu
  • E. R. Hancock
  • W. A. P. Smith
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5654)

Abstract

We generalize the classic self-organizing map (SOM) in flat Euclidean space (linear manifold) onto a Riemannian manifold. Both sequential and batch learning algorithms for the generalized SOM are presented. Compared with the classical SOM, the most novel feature of the generalized SOM is that it can learn the intrinsic topological neighborhood structure of the underlying Riemannian manifold that fits to the input data. We here compared the performance of the generalized SOM and the classical SOM using real 3-Dimensional facial surface normals data. Experimental results show that the generalized SOM outperforms the classical SOM when the data lie on a curved Riemannian manifold.

Keywords

SOM Riemannian Manifold Manifold Learning Pattern Recognition 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Kohonen, T.: Self-Organization and Associative Memory, 2nd edn. Springer, Berlin (1988)CrossRefMATHGoogle Scholar
  2. 2.
    Kohonen, T., Kaski, S., Lagus, K., Salojarvi, J., Honkela, J., Paatero, V., Saarela, A.: Self Organization of a Massive Document Collection. IEEE Transactions on Neural Networks 11(3), 574–585 (2000)CrossRefGoogle Scholar
  3. 3.
    Mavroudi, S., Papadimitriou, S., Bezerianos, A.: Gene Expression Data Analysis with a Dynamically Extended Self-organized Map That Exploits Class Information. Bioinformatics 18(11), 1446–1453 (2000)CrossRefGoogle Scholar
  4. 4.
    Joutsiniemi, S.L., Kaski, S., Larsen, T.A.: Self-organizing Map in Recognition of Topographic Patterns of EEG Spectra. Bioinformatics 42(11), 1062–1068 (1995)Google Scholar
  5. 5.
    Laaksonen, J., Koskela, M., Oja, E.: Class Distributions on SOM Surfaces for Feature Extraction and Object Retrieval. Neural Networks 17(8), 1121–1133 (2004)CrossRefGoogle Scholar
  6. 6.
    Barreto, G.A., Arajo, A.F.R.: Identification and Control of Dynamical Using the Self-organizing Map. IEEE Transactions on Neural Networks 15(5), 1244–1259 (2004)CrossRefGoogle Scholar
  7. 7.
    Wu, S.T., Chow, T.W.S.: PRSOM: A New Visualization Method by Hybridizing Multidimensional Scaling and Self-organizing Map. IEEE Transactions on Neural Networks 16(6), 1362–1380 (2005)CrossRefGoogle Scholar
  8. 8.
    Andras, P.: Kernel-Kohonen Networks. Int. J. Neural. Syst. 12(2), 117–135 (2002)CrossRefGoogle Scholar
  9. 9.
    Manuel, A.M., Alberto, M.: Extending the SOM algorithm to Non-Euclidean Distances via the Kernel Trick. In: 11th International Conference on Neural-Information-Processing, pp. 150–157. Springer Press, Calcutta (2004)Google Scholar
  10. 10.
    Ritter, H.: Self-organiring Maps on non-Euclidean Spaces. In: WSOM 1999, pp. 1321–1344. IEEE Press, Espoo (1999)Google Scholar
  11. 11.
    Shi, C.Q., Zhang, S.L., Shi, Z.Z.: Geodesic Distance based SOM for Image Clustering. In: International Conference on Sensing, Computing and Automation, pp. 2483–2488. Watam Press, Chongqing (2006)Google Scholar
  12. 12.
    Simila, T.: Self-organizing Map Learning Nonlinearly Embedded Manifolds. J. Information Visualization 4, 22–31 (2005)CrossRefGoogle Scholar
  13. 13.
    Pennec, X.: Probabilities and Statistics on Riemannian Manifolds: Basic Tools for Geometric Measurements. In: IEEE Workshop on Nonlinear Signal and Image Processing, pp. 194–198. IEEE Press, Antalya (1999)Google Scholar
  14. 14.
    Karcher, H.: Riemannian Center of Mass and Mollifier Smoothing. Commun. Pure Appl. Math. 30(5), 509–541 (1977)MathSciNetCrossRefMATHGoogle Scholar
  15. 15.
    Jost, J.: Riemannian Geometry and Geometric Analysis, 3rd edn. Springer, Berlin (2002)CrossRefMATHGoogle Scholar
  16. 16.
    Fletcher, P.T., Lu, C., Pizer, S.M., Joshi, S.: Principal Geodesic Analysis for The Study of Nonlinear Statistics of Shape. IEEE Transactions on Medical Imaging 23(8), 995–1005 (2004)CrossRefGoogle Scholar
  17. 17.
    Chang, K., Bowyer, K.W., Flynn, P.J.: Face Recognition using 2D and 3D Facial Data. In: ACM Workshop on Multimodal User Authentication, pp. 25–32. ACM Press, Santa Barbara (2003)Google Scholar
  18. 18.
    Smith, W.A.P., Hancock, E.R.: A New Framework for Grayscale and Colour Non-lambertian Shape-from-Shading. In: Yagi, Y., Kang, S.B., Kweon, I.S., Zha, H. (eds.) ACCV 2007, Part II. LNCS, vol. 4844, pp. 869–880. Springer, Heidelberg (2007)CrossRefGoogle Scholar
  19. 19.
    Wu, J., Smith, W.A.P., Hancock, E.R.: Gender Classification using Shape from Shading. In: BMVC, UK, pp. 499–508 (2007)Google Scholar
  20. 20.
    Phillips, P.J., Moon, H., Rauss, P.J., Rizvi, S.: The FERET Evaluation Methodology for Face Recognition Algorithms. Transactions on Pattern Analysis and Machine Intelligence 22(10), 1090–1104 (2000)CrossRefGoogle Scholar
  21. 21.
    Tan, X.Y., Chen, S.C., Zhou, Z.H., Zhang, F.Y.: Recognizing Partially Occluded, Expression Variant Faces from Single Training Image per Person with SOM and Soft kNN Ensemble. IEEE Transactions on Neural Networks 16(4), 875–886 (2005)CrossRefGoogle Scholar
  22. 22.
    Ruggeri, M.R., Darom, T., Saupe, D., Kiryati, N.: Approximating Geodesics on Point Set Surface. In: EG/IEEE Symposium on Point-Based Graphics, pp. 85–93. IEEE Press, Los Alamitos (2006)Google Scholar
  23. 23.
    Stawiaski, J., Decencifre, E., Bidault, F.: Computing Approximate Geodesics and Minimal Surfaces using Watershed and Graph-cuts. In: 8th International Symposium on Mathematical Morphology, Brazil, pp. 349–360 (2007)Google Scholar
  24. 24.
    Surazhsky, V., Surazhsky, T., Kirsanov, D., Gortler, S.J., Hoppe, H.: Fast Exact and Approximate Geodesics on Meshes. ACM Transactions on Graphics 24(3), 553–560 (2005)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2009

Authors and Affiliations

  • D. J. Yu
    • 1
    • 2
  • E. R. Hancock
    • 2
  • W. A. P. Smith
    • 2
  1. 1.School of Computer ScienceNanjing University of Science and TechnologyChina
  2. 2.Department of Computer ScienceUniversity of YorkUK

Personalised recommendations