Adaptive Computer Vision: Online Learning for Object Recognition

  • Holger Bekel
  • Ingo Bax
  • Gunther Heidemann
  • Helge Ritter
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3175)

Abstract

The “life” of most neural vision systems splits into a one-time training phase and an application phase during which knowledge is no longer acquired. This is both technically inflexible and cognitively unsatisfying. Here we propose an appearance based vision system for object recognition which can be adapted online, both to acquire visual knowledge about new objects and to correct erroneous classification. The system works in an office scenario, acquisition of object knowledge is triggered by hand gestures. The neural classifier offers two ways of training: Firstly, the new samples can be added immediately to the classifier to obtain a running system at once, though at the cost of reduced classification performance. Secondly, a parallel processing branch adapts the classification system thoroughly to the enlarged image domain and loads the new classifier to the running system when ready.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Heidemann, G., Rae, R., Bekel, H., Bax, I., Ritter, H.: Integrating context-free and context-dependent attentional mechanisms for gestural object reference. In: Proc. Int’l Conf. Cognitive Vision Systems, Graz, Austria, pp. 22–33 (2003)Google Scholar
  2. 2.
    Heidemann, G., Ritter, H.: Efficient Vector Quantization Using the WTA-rule with Activity Equalization. Neural Processing Letters 13(1), 17–30 (2001)MATHCrossRefGoogle Scholar
  3. 3.
    Kahn, R.E., Swain, M.J., Prokopowicz, P.N., Firby, R.J.: Gesture recognition using the perseus architecture. Technical Report TR-96-04, 19 (1996)Google Scholar
  4. 4.
    Kalinke, T., von Seelen, W.: Entropie als Maß des lokalen Informationsgehalts in Bildern zur Realisierung einer Aufmerksamkeitssteuerung. In: Jähne, B., Geißler, P., Haußecker, H., Hering, F. (eds.) Mustererkennung 1996, pp. 627–634. Springer, Heidelberg (1996)Google Scholar
  5. 5.
    Kohonen, T.: Self-Organizing Maps. Springer, Heidelberg (1995)Google Scholar
  6. 6.
    Leonardis, A., Bischof, H., Maver, J.: Multiple eigenspaces. Pattern Recognition 35(11), 2613–2627 (2002)MATHCrossRefGoogle Scholar
  7. 7.
    Andersen, H.J., Stoerring, M., Granum, E.: Physics-based modelling of human Skin colour under mixed illuminants. Robotics and Autonomous Systems 35(3-4), 131–142 (2001)MATHCrossRefGoogle Scholar
  8. 8.
    Mel, B.W.: SEEMORE: Combining color, shape, and texture histogramming in a neurally-inspired approach to visual object recognition. Neural Computation 9, 777–804 (1997)CrossRefGoogle Scholar
  9. 9.
    Murase, H., Nayar, S.K.: Visual Learning and Recognition of 3-D Objects from Appearance. Int’l J. of Computer Vision 14, 5–24 (1995)CrossRefGoogle Scholar
  10. 10.
    Ossola, J.C., Bremond, F., Thonnat, M.: A communication level in a distributed architecture for object recognition. In: 8th International Conference on Systems Research Informatics and Cybernetics (August 1996)Google Scholar
  11. 11.
    Ritter, H.J., Martinetz, T.M., Schulten, K.J.: Neuronale Netze. Addison-Wesley, München (1992)MATHGoogle Scholar
  12. 12.
    Sanger, T.D.: Optimal Unsupervised Learning in a Single-Layer Linear Feedforward Neural Network. Neural Networks 2, 459–473 (1989)CrossRefGoogle Scholar
  13. 13.
    Theis, C., Iossifidis, I., Steinhage, A.: Image Processing Methods for Interactive Robot Control. In: Proc. IEEE Roman International Workshop on Robot-Human Interactive Communication, Bordeaux and Paris, France (2001)Google Scholar
  14. 14.
    Tipping, M.E., Bishop, C.M.: Mixtures of probabilistic principal component analyzers. Neural Computation 11(2), 443–482 (1999)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Holger Bekel
    • 1
  • Ingo Bax
    • 1
  • Gunther Heidemann
    • 1
  • Helge Ritter
    • 1
  1. 1.AG NeuroinformaticsBielefeld UniversityBielefeldGermany

Personalised recommendations