Small Sample Learning of Superpixel Classifiers for EM Segmentation

  • Toufiq Parag
  • Stephen Plaza
  • Louis Scheffer
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8673)


Pixel and superpixel classifiers have become essential tools for EM segmentation algorithms. Training these classifiers remains a major bottleneck primarily due to the requirement of completely annotating the dataset which is tedious, error-prone and costly. In this paper, we propose an interactive learning scheme for the superpixel classifier for EM segmentation. Our algorithm is ‘active semi-supervised’  because it requests the labels of a small number of examples from user and applies label propagation technique to generate these queries. Using only a small set (< 20%) of all datapoints, the proposed algorithm consistently generates a classifier almost as accurate as that estimated from a complete groundtruth. We provide segmentation results on multiple datasets to show the strength of these classifiers.


Label Propagation Active Learning Algorithm arXiv Version Label Propagation Method Motion Detection Circuit 
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.


  1. 1.
    Takemura, S.Y., et al.: A visual motion detection circuit suggested by Drosophila connectomics. Nature 500(7461), 175–181 (2013)Google Scholar
  2. 2.
    Arbelaez, P., Maire, M., Fowlkes, C., Malik, J.: Contour detection and hierarchical image segmentation. IEEE Transactions on PAMI 33(5), 898–916 (2011)Google Scholar
  3. 3.
    Funke, J., Andres, B., Hamprecht, F., Cardona, A., Cook, M.: Efficient automatic 3D-reconstruction of branching neurons from EM data. In: CVPR (2012)Google Scholar
  4. 4.
    Kaynig, V., Fuchs, T., Buhmann, J.: Neuron geometry extraction by perceptual grouping in sstem images. In: CVPR (2010)Google Scholar
  5. 5.
    Vitaladevuni, S., Basri, R.: Co-clustering of image segments using convex optimization applied to em neuronal reconstruction. In: CVPR (2010)Google Scholar
  6. 6.
    Chklovskii, D.B., Vitaladevuni, S., Scheffer, L.K.: Semi-automated reconstruction of neural circuits using electron microscopy. Current Opinion in Neurobiology 20(5), 667–675 (2010)CrossRefGoogle Scholar
  7. 7.
    Jain, V., Turaga, S.C., Briggman, K., Helmstaedter, M.N., Denk, W., Seung, H.S.: Learning to agglomerate superpixel hierarchies. In: NIPS, vol. 24, pp. 648–656 (2011)Google Scholar
  8. 8.
    Andres, B., Köthe, U., Helmstaedter, M., Denk, W., Hamprecht, F.: Segmentation of SBFSEM Volume Data of Neural Tissue by Hierarchical Classification. Pattern Recognition 5096(15), 142–152 (2008)CrossRefGoogle Scholar
  9. 9.
    Andres, B., Kroeger, T., Briggman, K.L., Denk, W., Korogod, N., Knott, G., Koethe, U., Hamprecht, F.A.: Globally optimal closed-surface segmentation for connectomics. In: Fitzgibbon, A., Lazebnik, S., Perona, P., Sato, Y., Schmid, C. (eds.) ECCV 2012, Part III. LNCS, vol. 7574, pp. 778–791. Springer, Heidelberg (2012)CrossRefGoogle Scholar
  10. 10.
    Jain, V., Bollmann, B., Richardson, M., Berger, D., Helmstaedter, M., Briggman, K., Denk, W., Bowden, J., Mendenhall, J., Abraham, W., Harris, K., Kasthuri, N., Hayworth, K., Schalek, R., Tapia, J., Lichtman, J., Seung, H.: Boundary learning by optimization with topological constraints. In: CVPR (2010)Google Scholar
  11. 11.
    Beucher, S., Meyer, F.: The Morphological Approach to Segmentation: The Watershed Transformation. In: Mathematical Morphology in Image Processing, pp. 433–481 (1993)Google Scholar
  12. 12.
    Nunez-Iglesias, J., Kennedy, R., Parag, T., Shi, J., Chklovskii, D.B.: Machine learning of hierarchical clustering to segment 2D and 3D images. PLoS ONE 8(8) (August 2013)Google Scholar
  13. 13.
    Helmstaedter, M.: Cellular-resolution connectomics: challenges of dense neural circuit reconstruction. Nat. Methods 10(6), 501–507 (2013)CrossRefGoogle Scholar
  14. 14.
    Sommer, C., Straehle, C., Koethe, U., Hamprecht, F.A.: Ilastik: Interactive learning and segmentation toolkit. In: ISBI (2011)Google Scholar
  15. 15.
    Parag, T., Chakraborty, A., Plaza, S.: A context-aware delayed agglomeration framework for EM segmentation. arXiv 1406:1476 (2014)Google Scholar
  16. 16.
    Beygelzimer, A., Dasgupta, S., Langford, J.: Importance weighted active learning. In: ICML 2009 (2009)Google Scholar
  17. 17.
    Zhu, X., Lafferty, J., Ghahramani, Z.: Combining Active Learning and Semi-Supervised Learning Using Gaussian Fields and Harmonic Functions. In: ICML 2003 Workshop on The Continuum from Labeled to Unlabeled Data in Machine Learning and Data Mining (2003)Google Scholar
  18. 18.
    Muslea, I., Minton, S., Knoblock, C.A.: Active + semi-supervised learning = robust multi-view learning. In: ICML (2002)Google Scholar
  19. 19.
    Chapelle, O., Schölkopf, B., Zien, A. (eds.): Semi-Supervised Learning. MIT Press, Cambridge (2006)Google Scholar
  20. 20.
    Breiman, L.: Random forests. Machine Learning 45(1), 5–32 (2001)CrossRefzbMATHGoogle Scholar
  21. 21.
    Spielman, D.A., Teng, S.H.: A local clustering algorithm for massive graphs and its application to nearly-linear time graph partitioning. CoRR abs/0809.3232 (2008)Google Scholar
  22. 22.
    Koutis, I., Miller, G., Peng, R.: A nearly-m log n time solver for sdd linear systems. In: Foundations of Computer Science (FOCS), pp. 590–598 (2011)Google Scholar
  23. 23.
    Demidov, D.: Algebraic multigrid solver,
  24. 24.
    Zhu, X., Goldberg, A.B., Brachman, R., Dietterich, T.: Introduction to Semi-Supervised Learning. Morgan and Claypool Publishers (2009)Google Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Toufiq Parag
    • 1
  • Stephen Plaza
    • 1
  • Louis Scheffer
    • 1
  1. 1.Janelia Farm Research Campus- HHMIAshburnUSA

Personalised recommendations