Optimization methods in multilayer classifier networks for automatic control of lamellibranch larva growth

  • György G. Vass
  • Mohamed Daoudi
  • Faouzi Ghorbel
Poster Session C: Compression, Hardware & Software, Image Database, Neural Networks, Object Recognition & Reconstruction
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1311)

Abstract

The problem considered here is the age discrimination of lamellibranch larvae. Patterns of larvae are presented to a multilayer feedforward neural network. Samples are represented by shape descriptors calculated on the basis of a normalized arc length parametrization of their boundary. After training, the network will classify samples on the basis of their characteristic shapes. In neural network applications one often faces the problem of optimal network size, which is an implicit function of problem complexity and available amount of data for training. This paper presents some possible solutions to cope with this problem. Results obtained are compared with previous experiments on feedforward networks.

Keywords

Hide Neuron Shape Descriptor Fourier Descriptor Scatter Matrice Multilayer Perceptron Neural Network 
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.

References

  1. [1]
    Webb, A.R., Lowe, D.: “The Optimised Internal Representation of Multilayer Classifier Networks Performs Nonlinear Discriminant Analysis”, Neural Networks, Vol. 3, pp. 367–375 (1990)CrossRefGoogle Scholar
  2. [2]
    Rees, C.B.: “The identification and classification of lamellibranch larvae”, Hull. Bull. Mar. Ecol. Vol. 3, No. 19, 73–104 (1950)Google Scholar
  3. [3]
    Ghorbel, F.,de Bougrenet de la Tocnaye, J.L.: “Automatic Control of Lamellibranch Larva Growth Using Contour Invariant Feature Extraction”, Pattern Recognition, Vol.23, No. 3 (1990)Google Scholar
  4. [4]
    Vass Gy., Daoudi, M.: “Automatic Control of Lamellibranch Larva Growth using Neural Network Classifier” Proc. of CESA'96, Lille, France, pp. 531-536, July (1996)Google Scholar
  5. [5]
    Bez, H.E.: “On analysis of symmetry for plane curves and plane curve algorithms”, Comp. Aided Geom. Design, Vol.9, pp. 125–142 (1992)CrossRefGoogle Scholar
  6. [6]
    Ghorbel, F., Burdin, V.: “Invariant Approximation of Star-Shaped Form for Medical Applications”, Curves and Surfaces II, P.J. Laurent, A. Le Méhauté and L. L. Schumaker (eds.), AKPeters, Boston, pp. 1–9 (1991)Google Scholar
  7. [7]
    Mao, J.: “Design and Analysis of Neural Networks for Pattern Recognition”, Dissertation submitted to Michigan State University (1994)Google Scholar
  8. [8]
    Pao, Y.H.: “Adaptive Pattern Recognition and Neural Networks”, AddisonWesley, Reading, Massachusetts (1989)Google Scholar
  9. [9]
    Lengellé, R., Denoeux, T.: “Training MLPs Layer by Layer Using an Objective Function for Internal Representations”, Neural Networks, Vol. 9, No. 1, pp. 83–97 (1996)CrossRefGoogle Scholar
  10. [10]
    Gallinari, P., et al.: “On the Relations Between Discriminant Analysis and Multilayer Perceptrons”, Neural Networks, Vol. 4, pp. 349–360 (1991)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • György G. Vass
    • 1
  • Mohamed Daoudi
    • 2
  • Faouzi Ghorbel
    • 3
  1. 1.Department ofMicrowave TelecommunicationsBudapest University of TechnologyBudapestHongrie
  2. 2.Département Informatique et RéseauxRéseaux
  3. 3.Groupe de Recherche Images et Formes ENIC/INTEcole Nouvelle d'Ingénieurs en Communication Rue Guglielmo MarconiVilleneuve d'Ascq Cedex

Personalised recommendations