Image compression using feedforward neural networks — Hierarchical approach

  • Stanislaw Osowski
  • Robert Waszczuk
  • Piotr Bojarczak
Neural Networks for Perception
Part of the Lecture Notes in Computer Science book series (LNCS, volume 930)


The paper presents the hierarchical approach to the problem of image compression using feedforward neural networks. In this approach smaller frames are used in the regions containing more details and larger, when the grey level is uniform. Thanks to this the number of data is reduced, learning speed accelerated and the quality of compression improved. The numerical results, confirming the efficiency of the proposed approach and good generalization properties are presented and discussed.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    J. Hertz, A. Krogh, R. Palmer, Introduction to the Theory of Neural Computation, Addison Wesley, 1991, AmsterdamGoogle Scholar
  2. [2]
    Cottrell G., Munro P., and Zipser D., Image compression by back propagation: an example of extensional programming, Technical Report ICS report 8702, ICS-UCSD, San Diego, California, USA, February 1987.Google Scholar
  3. [3]
    Kunt M., Image compression using neural networks, Technical Report EPFL Lausanne, (CARNAC), 1993Google Scholar
  4. [4]
    Mougeot M., Azencott R., Angeniol B., Image compression with backpropagation: improvement of the visual restoration using different cost functions, Neural Networks, 1991, vol. 4, pp. 467–476Google Scholar
  5. [5]
    Oja E., Wang L., Image compression by MLP and PCA neural networks, SCIA-93, Tromso, 1993, pp. 1317–1324Google Scholar
  6. [6]
    Sonehara N., Kawato M., Miyake S., Nakane K., Image data compression using neural networks, IJCNN, Washington, 1989, pp. II35–II40Google Scholar
  7. [7]
    G. Vines, M. Hayes III, Map search strategies for IFS image compression algorithms, 14 Colloque Gretsi, Juan-Les-Pins, 1993, pp. 843–846Google Scholar
  8. [8]
    G. Martinelli, L. Prina Ricotti, G. Marcone, Neural ckustering for optimal KLT image compression, IEE Trans. Signal Processing, 1993, vol. 41, pp. 1737–1739Google Scholar
  9. [9]
    P. Yu, A. N. Vanetsopoulos, Hierarchical multirate vector quantization for image coding, Signal Processing: Image communication, 1992, vol. 4, pp. 497–505Google Scholar
  10. [10]
    S. Osowski, J. Herault, Signal flow graphs as an efficient tool for exact gradient and hessian determination, Complex Systems (accepted for publication)Google Scholar
  11. [11]
    S. Osowski, Fast learning algorithms for feedforward multilayer neural networks, IEEE 1994 IMACS Int. Symp. Signal Processing and Neural Networks, Lille, 1994, pp. 27–30Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1995

Authors and Affiliations

  • Stanislaw Osowski
    • 1
  • Robert Waszczuk
    • 1
  • Piotr Bojarczak
    • 1
  1. 1.Institute of the Theory of Electrical Engineering and Electrical MeasurementsTechnical University 00-661 WarsawPoland

Personalised recommendations