Advertisement

Neural Networks versus Image Pyramids

  • Horst Bischof
  • Walter G. Kropatsch

Abstract

Neural networks and image pyramids are massively parallel processing structures. In this paper we exploit the similarities as well as the differences between these structures. The general goal is to exchange knowledge between these two fields. After introducing the basic concepts of neural networks and image pyramids we give a translation table of the vocabulary used in image pyramids and those used in neural networks. Image pyramids which store and process numerical information (e.g. grey values of pixels) are very similar to neural networks. Therefore we concentrate on ”symbolic pyramids”. The main idea is to replace a cell of the pyramid by a small neural network, in order to represent and process symbolic information. We will consider local as well as distributed representations for symbolic information. In particular we present a neural implementation of the 2 x 2/2 curve pyramid. We derive some general rules for implementing symbolic pyramids by neural networks. Finally we briefly discuss the role of learning in image pyramids.

Keywords

Neural Network Transitive Closure Neighborhood Graph Symbolic Information Curve Relation 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    Horst Bischof. Modular, hierarchical, and geometrical neural networks. Technical Report PRIP-TR-9, Dept. for Pattern Recognition and Image Processing, TU Wien, 1991.Google Scholar
  2. [2]
    Horst Bischof. Neural networks and image pyramids. In Bischof and Kropatsch [3], pages 249-260.Google Scholar
  3. [3]
    Horst Bischof and Walter Kropatsch, editors. Pattern Recognition 1992, volume 62 of OCG-Schriftenreihe. Oldenbourg, 1992.Google Scholar
  4. [4]
    Horst Bischof and Walter G. Kropatsch. Neural Networks versus Image Pyramids. Technical Report PRIP-TR-7, Dept. for Pattern Recognition and Image Processing, TU Wien, 1993.Google Scholar
  5. [5]
    Horst Bischof and Axel Pinz. The invariance problem for hierarchical neural networks. In Chen Su-Shing, editor, Neural and Stochastic Methods in Image and Signal Processing, volume SPIE Vol. 1766, pages 118–129. SPIE, 1992.Google Scholar
  6. [6]
    M. Bister, J. Cornelis, and Azriel Rosen-feld. A critical view of pyramid segmentation algorithms. Pattern Recognition Letters, Vol. ll (No. 9): pp. 605–617, September 1990.CrossRefGoogle Scholar
  7. [7]
    J.A. Feldman. Connectionist models and their properties. Cognitive Science, 6: 205–254, 1982.CrossRefGoogle Scholar
  8. [8]
    J.A. Fodor and Z.W. Pylyshyn. Connectionism and cognitive architecture: A critical analysis. Cognition, 28: 3–71, 1988.CrossRefGoogle Scholar
  9. [9]
    K. Fukushima, S. Miyake, and T. Ito. Neocognitron: a neural network model for a mechanism of visual pattern recognition. IEEE Transactions on Systems, Man, and Cybernetics, Vol. SMC-13(Nb. 3):pp. 826–834, September/October 1983.Google Scholar
  10. [10]
    William I. Grosky and Ramesh Jain. A pyramidbased approach to segmentation applied to region matching. IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI-8(No.5): pp. 639–650, September 1986.CrossRefGoogle Scholar
  11. [11]
    P.J. Hancock. Data representation in neural nets: an empirical study. In Proc. of the 1988 Connectionist Models Summer School, pages 11-20. Morgan Kaufmann, 1988.Google Scholar
  12. [12]
    Robert M. Haralick and Shapiro Linda G. Glossary of computer vision terms. Pattern Recognition, 24: pp. 69–93, 1991.CrossRefGoogle Scholar
  13. [13]
    R. L. Hartley. Multi-Scale Models in Image Analysis. PhD thesis, University of Maryland, Computer Science Center, 1984.Google Scholar
  14. [14]
    D.O. Hebb. The organization of behavior, Wiley, New York, 1949.Google Scholar
  15. [15]
    G. Hinton. Connectionist Learning Procedures. Artificial Intelligence, 40: 185–234, 1989.CrossRefGoogle Scholar
  16. [16]
    G.E. Hinton, J.L. McCelland, and D.E. Rumelhart. Distributed Representations. In McCelland Rumelhart, editor, Parallel Distributed Processing, volume 1. MIT Press, 1986.Google Scholar
  17. [17]
    G.E. Hinton and T.J. Sejnowski. Learning and relearning in boltzmann machines. In McCelland Rumelhart, editor, Parallel Distributed Processing, volume 1. MIT Press, 1986.Google Scholar
  18. [18]
    J.E. Hopcroft and Ullman J.D. Introduction to Automata theory, languages and computations. Addison Wesley, 1979.Google Scholar
  19. [19]
    J.J. Hopfield. Neural networks and physical systems with emergent collective computational abilities. PNAS, 79: 2554–2558, 1982.MathSciNetCrossRefGoogle Scholar
  20. [20]
    J.J. Hopfield. Neurons with graded response have collective computational properties like those of two-state neurons. PNAS, 82: 3088–3092, 1984.CrossRefGoogle Scholar
  21. [21]
    Walter G. Kropatsch. Hierarchical curve representation in a new pyramid scheme. Technical Report TR-1522, University of Maryland, Computer Science Center, June 1985.Google Scholar
  22. [22]
    Walter G. Kropatsch. A pyramid that grows by powers of 2. Pattern Recognition Letters, Vol. 3: pp. 315–322, 1985.CrossRefGoogle Scholar
  23. [23]
    Walter G. Kropatsch. Curve representations in multiple resolutions. In Proc. Eighth International Conference on Pattern Recognition, pages 1283-1285. IEEE Comp.Soc., 1986.Google Scholar
  24. [24]
    Walter G. Kropatsch. Elimination von ”kleinen” Kurvenstücken in der 2 x 2/2 Kurvenpyramide. In E. Paulus, editor, Mustererkennung 1987, Informatik Fachberichte 149, pages 156-160. Springer Verlag, 1987.Google Scholar
  25. [25]
    Y. Le Cun, O. Matan, Boser B., Denker J.S., Henderson D., Howard R.E., Hubbard W., Jackel L.D., and Baird H.S. Handwritten ZIP Code Recognition with Multilayer Networks. In Proc. of the 10.ICPR, pages 35-40. IEEE Computer Society, 1990.Google Scholar
  26. [26]
    S. Levialdi. Programming image processing machines. In S. Levialdi and V. Cantoni, editors, Pyramidal Systems for Image Processing and Computer Vision, volume F25 of NATO ASI Series, pages 311–328. Springer-Verlag Berlin, Heidelberg, 1986.CrossRefGoogle Scholar
  27. [27]
    R. Linsker. Self-organization in a perceptual network. IEEE Computer, 21: 105–117, 1988.CrossRefGoogle Scholar
  28. [28]
    Peter Meer. Stochastic image pyramids. Computer Vision, Graphics, and Image Processing, Vol. 45 (No. 3):pp. 269–294, March 1989.CrossRefGoogle Scholar
  29. [29]
    Jordan M.I. Attractor dynamics and parallelism in a connectionist sequential machine. In Proc. of 8th Annual Conf. of the Cognitive Science Society, pages 531-546, 1986.Google Scholar
  30. [30]
    A. Montanvert and Bertolino P. Irregular pyramids for parallel image segmentation. In Bischof and Kropatsch [3], pages 13-34.Google Scholar
  31. [31]
    E. Oja. A simplified neuron model as a priniple component analyzer. J. Mathematical Biology, 15: 267–273, 1982.MathSciNetMATHCrossRefGoogle Scholar
  32. [32]
    Smolensky Paul. Tensor product variable binding and the representation of symbolic structures in connectionist systems. Artificial Intelligence, 46: 159–216, 1990.MathSciNetMATHCrossRefGoogle Scholar
  33. [33]
    J.B. Pollack. Recursive distributed representations. Artificial Intelligence, 46: 77–105, 1990.CrossRefGoogle Scholar
  34. [34]
    Azriel Rosenfeld, editor. Multiresolution Image Processing and Analysis. Springer, Berlin, 1984.MATHGoogle Scholar
  35. [35]
    Azriel Rosenfeld. Arc colorings, partial path groups, and parallel graph contractions. Technical Report TR-1524, University of Maryland, Computer Science Center, July 1985.Google Scholar
  36. [36]
    David E. Rumelhart and James A. McClelland. Parallel Distributed Processing, volume 1. MIT Press, first edition, 1986.Google Scholar
  37. [37]
    Steven L. Tanimoto. Paradigms for pyramid machine algorithms. In S. Levialdi and V. Cantoni, editors, Pyramidal Systems for Image Processing and Computer Vision, volume F25 of NATO ASI Series, pages 173–194. Springer-Verlag Berlin, Heidelberg, 1986.CrossRefGoogle Scholar
  38. [38]
    Toru Yamaguchi and Walter G. Kropatsch. A vision by neural network or by pyramid. In Proceedings of the 6th Scandinavian Conference on Image Analysis, pages 104–111, Oulu, Finland, June 1989.Google Scholar

Copyright information

© Springer-Verlag/Wien 1993

Authors and Affiliations

  • Horst Bischof
    • 1
  • Walter G. Kropatsch
    • 1
  1. 1.Inst. for Automation Dept. for Pattern Recognition and Image ProcessingTechnical University ViennaAustria

Personalised recommendations