Segmentation of SBFSEM Volume Data of Neural Tissue by Hierarchical Classification

  • Björn Andres
  • Ullrich Köthe
  • Moritz Helmstaedter
  • Winfried Denk
  • Fred A. Hamprecht
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5096)


Three-dimensional electron-microscopic image stacks with almost isotropic resolution allow, for the first time, to determine the complete connection matrix of parts of the brain. In spite of major advances in staining, correct segmentation of these stacks remains challenging, because very few local mistakes can lead to severe global errors. We propose a hierarchical segmentation procedure based on statistical learning and topology-preserving grouping. Edge probability maps are computed by a random forest classifier (trained on hand-labeled data) and partitioned into supervoxels by the watershed transform. Over-segmentation is then resolved by another random forest. Careful validation shows that the results of our algorithm are close to human labelings.


Random Forest Convolutional Neural Network Initial Segmentation Isotropic Resolution Watershed Algorithm 
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.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Briggman, K.L., Denk, W.: Towards neural circuit reconstruction with volume electron microscopy techniques. Curr. Opinion in Neurobiology 16, 562–570 (2006)CrossRefGoogle Scholar
  2. 2.
    Denk, W., Horstmann, H.: Serial block-face scanning electron microscopy to reconstruct three-dimensional tissue nanostructure. PLoS Biology 2, 1900–1909 (2004)CrossRefGoogle Scholar
  3. 3.
    Kaynig, V., Fischer, B., Buhmann, J.M.: Probabilistic Image Registration and Anomaly Detection by Nonlinear Warping Technical Report. ETH Zürich (2008)Google Scholar
  4. 4.
    Meine, H., Köthe, U.: The GeoMap: A unified representation for topology and geometry. In: Brun, L., Vento, M. (eds.) GbRPR 2005. LNCS, vol. 3434, pp. 132–141. Springer, Heidelberg (2005)Google Scholar
  5. 5.
    Jain, V., Murray, J.F., Roth, F., Turaga, S., Zhigulin, V., Briggman, K.L., Helmstädter, M.N., Denk, W., Seung, H.S.: Supervised learning of image restoration with convolutional networks. In: Proceedings of the ICCV 2007 (2007)Google Scholar
  6. 6.
    Jurrus, E., Tasdizen, T., Koshevoy, P., Fletcher, P.T., Hardy, M., Chien, C.B., Denk, W., Whitaker, R.: Axon tracking in serial block-face scanning electron microscopy. WS Med. Image Comp. and Computer-Assisted Intervention (2006)Google Scholar
  7. 7.
    Macke, J.H., Maack, N., Gupta, R., Denk, W., Schölkopf, B., Borst, A.: Contour-propagation algorithms for semi-automated reconstruction of neural processes. Journal of Neuroscience Methods 167(2), 349–357 (2008)CrossRefGoogle Scholar
  8. 8.
    Ren, X., Malik, J.: Learning a classification model for segmentation. In: Proceedings of the ICCV 2003, vol. 1, pp. 10–16 (2003)Google Scholar
  9. 9.
    Konishi, S., Yuille, A., Coughlan, J., Zhu, S.: Statistical edge detection: Learning and evaluating edge cues. PAMI 25(1), 57–74 (2003)Google Scholar
  10. 10.
    Martin, D.R., Fowlkes, C.C., Malik, J.: Learning to detect natural image boundaries using local brightness, color, and texture cues. PAMI 26(5), 530–549 (2004)Google Scholar
  11. 11.
    Dollar, P., Tu, Z., Belongie, S.: Supervised learning of edges and object boundaries. In: Proceedings of the CVPR 2006, vol. 2, pp. 1964–1971 (2006)Google Scholar
  12. 12.
    Derivaux, S., Lefevre, S., Wemmert, C., Korczak, J.: On machine learning in watershed segmentation. IEEE WS Mach. Learning for Signal Proc., 187–192 (2007)Google Scholar
  13. 13.
    Levner, I., Zhang, H.: Classification-driven watershed segmentation. IEEE Transactions on Image Processing 16, 1437–1445 (2007)CrossRefGoogle Scholar
  14. 14.
    Vanhamel, I., Pratikakis, I., Sahli, H.: Multiscale gradient watersheds of color images. IEEE Transactions on Image Processing 12, 617–626 (2003)CrossRefGoogle Scholar
  15. 15.
    Grimaud, M.: New measure of contrast: the dynamics. In: Gader, P.D., Dougherty, E.R., Serra, J.C. (eds.) Proc. Image Algebra and Morphological Image Processing III, vol. 1769, pp. 292–305. SPIE (1992)Google Scholar
  16. 16.
    de Bock, J., de Smet, P., Philips, W.: Watersheds and normalized cuts as basic tools for perceptual grouping. In: Proc. ProRISC 2004 (2004)Google Scholar
  17. 17.
    Stawiaski, J., Decenciere, E.: Region merging via graph-cuts. In: 12th International Congress for Stereology, ICS 2007 (2007)Google Scholar
  18. 18.
    Haris, K., Efstratiadis, S.N., Maglaveras, N., Katsaggelos, A.K.: Hybrid image segmentation using watersheds and fast region merging. IEEE Transactions on Image Processing 7, 1684–1699 (1998)CrossRefGoogle Scholar
  19. 19.
    Marcotegui, B., Beucher, S.: Fast implementation of waterfall based on graphs. In: Ronse, C., Najman, L., Decencière, E. (eds.) Proc. ISMM 2005, pp. 177–186 (2005)Google Scholar
  20. 20.
    Breiman, L.: Random forests. Machine Learning 45, 5–32 (2001)CrossRefzbMATHGoogle Scholar
  21. 21.
    Helmstaedter, M., Denk, W.: Forthcoming paper (2008)Google Scholar
  22. 22.
    Köthe, U.: Edge and junction detection with an improved structure tensor. In: Michaelis, B., Krell, G. (eds.) DAGM 2003. LNCS, vol. 2781, pp. 25–32. Springer, Heidelberg (2003)Google Scholar
  23. 23.
    Vincent, L., Soille, P.: Watersheds in digital spaces: an efficient algorithm based on immersion simulations. PAMI 13(6), 583–598 (1991)Google Scholar
  24. 24.
    Nguyen, H.T., Worring, M., van den Boomgaard, R.: Watersnakes: Energy-driven watershed segmentation. PAMI 25(3), 330–342 (2003)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Björn Andres
    • 1
  • Ullrich Köthe
    • 1
  • Moritz Helmstaedter
    • 2
  • Winfried Denk
    • 2
  • Fred A. Hamprecht
    • 1
  1. 1.Interdisciplinary Center for Scientific Computing (IWR)University of Heidelberg 
  2. 2.Max Planck Institute for Medical ResearchHeidelbergGermany

Personalised recommendations