Traffic Sign Classifier Adaption by Semi-supervised Co-training

  • Matthias Hillebrand
  • Ulrich Kreßel
  • Christian Wöhler
  • Franz Kummert
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7477)


The recognition of traffic signs in many state-of-the-art driver assistance systems is performed by statistical pattern classification methods. Traffic signs in European countries share many similarities but also vary in colour, size, and depicted symbols, making it hard to obtain one general classifier with good performance in all countries. Training separate classifiers for each country requires huge amounts of labelled training data. A well-trained classifier for one country can be adapted to other countries by semi-supervised learning methods to perform reasonably well with relatively low requirements regarding labelled training data. Self-training classifiers adapting themselves to unknown domains always risk that the adaption will become ineffective or even fail completely due to the occurrence of incorrectly labelled samples. To assure that self-training classifiers adapt themself correctly, advanced multi-classifier training methods like co-training are applied.


self-training semi-supervised co-training 


  1. 1.
    Blum, A., Mitchell, T.: Combining labeled and unlabeled data with co-training. In: Proceedings of the Eleventh Annual Conference on Computational Learning Theory, COLT 1998, pp. 92–100. ACM, New York (1998)CrossRefGoogle Scholar
  2. 2.
    Chang, C.C., Lin, C.J.: LIBSVM: A library for support vector machines. ACM Transactions on Intelligent Systems and Technology 2, 27:1–27:27 (2011)Google Scholar
  3. 3.
    Chapelle, O., Schölkopf, B., Zien, A. (eds.): Semi-Supervised Learning. Adaptive Computation and Machine Learning. The MIT Press (2006)Google Scholar
  4. 4.
    Cui, T., Grumpe, A., Hillebrand, M., Kreßel, U., Kummert, F., Wöhler, C.: Analytically tractable sample-specific confidence measures for semi-supervised learning. In: Proc. 21st Workshop Computational Intelligence, pp. 171–186 (2011)Google Scholar
  5. 5.
    Fu, M.Y., Huang, Y.S.: A survey of traffic sign recognition. In: 2010 International Conference on Wavelet Analysis and Pattern Recognition (ICWAPR), pp. 119–124 (2010)Google Scholar
  6. 6.
    Hillebrand, M., Wöhler, C., Krüger, L., Kreßel, U., Kummert, F.: Self-learning with confidence bands. In: Proc. 20th Workshop Computational Intelligence, pp. 302–313 (2010)Google Scholar
  7. 7.
    Hillebrand, M., Wöhler, C., Kreßel, U., Kummert, F.: Semi-supervised Training Set Adaption to Unknown Countries for Traffic Sign Classifiers. In: Schwenker, F., Trentin, E. (eds.) PSL 2011. LNCS, vol. 7081, pp. 120–127. Springer, Heidelberg (2012)CrossRefGoogle Scholar
  8. 8.
    Hoessler, H., Wöhler, C., Lindner, F., Kreßel, U.: Classifier training based on synthetically generated samples. In: The 5th International Conference on Computer Vision Systems (2007)Google Scholar
  9. 9.
    Lee, D.D., Seung, H.S.: Learning the parts of objects by non-negative matrix factorization. Letters to Nature 401(1), 788–791 (1999)Google Scholar
  10. 10.
    Lindner, F.: Adaptive Traffic Sign Recognition. Ph.D. thesis, Bielefeld University (2012)Google Scholar
  11. 11.
    Rokach, L.: Pattern Classification using Ensemble Methods. Series in Machine Perception and Artificial Intelligence, vol. 75. World Scientific (2010)Google Scholar
  12. 12.
    Schürmann, J.: Pattern Classification: A Unified View of Statistical and Neural Approaches. John Wiley & Sons, Inc. (1996)Google Scholar
  13. 13.
    Xu, Z., King, I., Lyu, M.R.: More Than Semi-supervised Learning. Lambert Academic Publishing (2010)Google Scholar
  14. 14.
    Zhu, X., Goldberg, A.B.: Introduction to Semi-Supervised Learning. Synthesis Lectures on Artificial Intelligence and Machine Learning. Morgan & Claypool Publishers (2009)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Matthias Hillebrand
    • 1
  • Ulrich Kreßel
    • 1
  • Christian Wöhler
    • 2
  • Franz Kummert
    • 3
  1. 1.Group Research and Advanced EngineeringDaimler AGUlmGermany
  2. 2.Image Analysis GroupTU DortmundDortmundGermany
  3. 3.Applied Informatics GroupBielefeld UniversityBielefeldGermany

Personalised recommendations