Do We Need Large Annotated Training Data for Detection Applications in Biomedical Imaging? A Case Study in Renal Glomeruli Detection

  • Michael GadermayrEmail author
  • Barbara Mara Klinkhammer
  • Peter Boor
  • Dorit Merhof
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10019)


Approaches for detecting regions of interest in biomedical image data mostly assume that a large amount of annotated training data is available. Certainly, for unchanging problem definitions, the acquisition of large annotated data is time consuming, yet feasible. However, distinct practical problems with large training corpi arise if variability due to different imaging conditions or inter-personal variations lead to significant changes in the image representation. To circumvent these issues, we investigate a classifier learning scenario which requires a small amount of positive annotation data only. Contrarily to previous approaches which focus on methodologies to explicitly or implicitly deal with specific classification scenarios (such as one-class classification), we show that existing supervised classification models can handle a changed setting effectively without any specific modifications.


Training Data Image Representation Negative Class Classifier Training Positive Class 
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.



This work was supported by the German Research Foundation (DFG), grant number ME3737/3-1.


  1. 1.
    Boor, P., et al.: Role of platelet-derived growth factor-CC in capillary rarefaction in renal fibrosis. Am. J. Pathol. 185, 2132–2142 (2015)CrossRefGoogle Scholar
  2. 2.
    Dauguet, J., Mangin, J.-F., Delzescaux, T., Frouin, V.: Robust inter-slice intensity normalization using histogram scale-space analysis. In: Barillot, C., Haynor, D.R., Hellier, P. (eds.) MICCAI 2004. LNCS, vol. 3216, pp. 242–249. Springer, Heidelberg (2004). doi: 10.1007/978-3-540-30135-6_30 CrossRefGoogle Scholar
  3. 3.
    Hegenbart, S., Uhl, A., Vécsei, A., Wimmer, G.: Scale invariant texture descriptors for classifying celiac disease. Med. Image Anal. 17(4), 458–474 (2013)CrossRefGoogle Scholar
  4. 4.
    Herve, N., Servais, A., Thervet, E., Olivo-Marin, J.C., Meas-Yedid, V.: Statistical color texture descriptors for histological images analysis. In: ISBI, pp. 724–727. IEEE press, New York (2011)Google Scholar
  5. 5.
    Jerman, T., Pernuš, F., Likar, B., Špiclin, Ž.: Computer-aided detection and quantification of intracranial aneurysms. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9350, pp. 3–10. Springer, Heidelberg (2015). doi: 10.1007/978-3-319-24571-3_1 CrossRefGoogle Scholar
  6. 6.
    Kato, T., Relator, R., Ngouv, H., Hirohashi, Y., Takaki, O., Kakimoto, T., Okada, K.: Segmental HOG: new descriptor for glomerulus detection in kidney microscopy image. BMC Bioinform. 16(1), 316 (2015)CrossRefGoogle Scholar
  7. 7.
    Kwitt, R., Hegenbart, S., Rasiwasia, N., Vécsei, A., Uhl, A.: Do we need annotation experts? A case study in celiac disease classification. In: Golland, P., Hata, N., Barillot, C., Hornegger, J., Howe, R. (eds.) MICCAI 2014. LNCS, vol. 8674, pp. 454–461. Springer, Heidelberg (2014). doi: 10.1007/978-3-319-10470-6_57 Google Scholar
  8. 8.
    Liu, W., Hua, G., Smith, J.R.: Unsupervised one-class learning for automatic outlier removal. In: CVPR, pp. 3826–3833. IEEE press, New York (2014)Google Scholar
  9. 9.
    Mirrashed, F., Rastegari, M.: Domain adaptive classification. In: ICCV, pp. 2608–2615. IEEE press, New York (2013)Google Scholar
  10. 10.
    Vahdat, A., Mori, G.: Handling uncertain tags in visual recognition. In: ICCV, pp. 737–744. IEEE press, New York (2013)Google Scholar
  11. 11.
    Xie, Y., Xing, F., Kong, X., Su, H., Yang, L.: Beyond classification: structured regression for robust cell detection using convolutional neural network. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 358–365. Springer, Heidelberg (2015). doi: 10.1007/978-3-319-24574-4_43 CrossRefGoogle Scholar
  12. 12.
    Zuluaga, M.A., Hush, D., Delgado Leyton, E.J.F., Hoyos, M.H., Orkisz, M.: Learning from only positive and unlabeled data to detect lesions in vascular CT images. In: Fichtinger, G., Martel, A., Peters, T. (eds.) MICCAI 2011. LNCS, vol. 6893, pp. 9–16. Springer, Heidelberg (2011). doi: 10.1007/978-3-642-23626-6_2 CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2016

Authors and Affiliations

  • Michael Gadermayr
    • 1
    Email author
  • Barbara Mara Klinkhammer
    • 2
  • Peter Boor
    • 2
  • Dorit Merhof
    • 1
  1. 1.Aachen Center for Biomedical Image Analysis, Visualization and Exploration (ACTIVE), Institute of Imaging and Computer VisionRWTH Aachen UniversityAachenGermany
  2. 2.Institute of PathologyUniversity Hospital Aachen, RWTH Aachen UniversityAachenGermany

Personalised recommendations