Neural Computing & Applications

, Volume 6, Issue 1, pp 2–11 | Cite as

Texture synthesis by a neural network model

  • B. B. Chaudhuri
  • P. Kundu


In this paper we propose a neural network model to synthesise texture images. The model is based on a continuous Hopfield-like network where each pixel of the image is occupied by a neuron that is eight-connected to its neighbours. A state of the neuron denotes a certain grey level of the corresponding pixel. The firing of the neuron changes its state, and hence the grey level of the corresponding pixel. Different two-tone and grey-tone texture images can be synthesised by manipulating the connection weights and by varying the algorithm iteration number. For grey-tone texture synthesis, a Markov chain principle has been employed to decide on the multiple state transition of a neuron. The model can be employed for texture propagation with the advantage that it allows propagation without showing any blocky effect.


Computer graphics Image processing Texture synthesis 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Fu KS, Lu SY. Stochastic tree grammer for texture synthesis and discrimination. Computer Graph Image Process 1979; 9: 234–245.Google Scholar
  2. 2.
    Yokoyama R, Haralick RM. Texture synthesis using a growth model. Computer Graph Image Process 1978; 8: 369–381.Google Scholar
  3. 3.
    Ahuja N. Mosaic models for image analysis and synthesis. PhD dissertation, Department of Computer Science, University of Maryland, 1979.Google Scholar
  4. 4.
    Schacter B, Rosenfeld A, Davis LS. Random mosaic models for textures. IEEE Trans System, Man and Cyber 1978; 8: 694–702.Google Scholar
  5. 5.
    Mezei L, Puzin M, Conroy P. Simulation of patterns of nature by computer graphics. Inform Process 1974; 74: 52–56.Google Scholar
  6. 6.
    Yokoyama R, Haralick RM. Texture pattern image generation by regular Markov chain. Pattern Recog 1979; 11: 225–234.Google Scholar
  7. 7.
    Chellappa R, Chatterjee S, Bagdazin R. Texture synthesis and compression using Gaussian-Markov random field models. IEEE Trans System, Man Cybern 1985; 15: 298–303.Google Scholar
  8. 8.
    Cross GR, Jain AK. Markov random field texture models. IEEE Trans Pattern Anal and Machine Intell 1983; 5: 1983.Google Scholar
  9. 9.
    Onural L, Gürelli MI. Generation and parameter estimation of Markov random field textures by highly parallel networks. From Pixels to Features 2, ESPRIT BRA Workshop on Parallelism in Image Process, Bonas, France, August 1990.Google Scholar
  10. 10.
    Mandlebrot BH. Fractals — Form, Chance, Dimension. W. H. Freeman, San Francisco, CA, 1977.Google Scholar
  11. 11.
    Voss RF. Fourier synthesis of Gaussian fractals: 1/f noises, landscapes and flakes. Tutorials on state of the art image synthesis. SIGGRAPH 1983, Detroit, MI, Vol. 10, 1983.Google Scholar
  12. 12.
    Johnson NL, Kotz S. Continuous Univariate Distribution — 2. Wiley.Google Scholar
  13. 13.
    Hopfield JJ. Neural networks and physical systems with emergent collective properties. Proc Nat Acad Sci USA 1982; 79: 2554–2558.Google Scholar
  14. 14.
    Hopfield JJ. Neurons with graded response have collective computational properties like those of two-state neurons. Proc Nat Acad Sci USA, Biophysics 1984; 81: 3088–3092.Google Scholar
  15. 15.
    Amin S, Gell M. Investigation of the Hopfield model and its attractors. Neural Comput and Applic 1994: 2: 129–133.Google Scholar
  16. 16.
    Kenemy J, Snell J. Finite Markov Chain. Springer, New York, 1960.Google Scholar

Copyright information

© Springer-Verlag London Limited 1997

Authors and Affiliations

  • B. B. Chaudhuri
    • 1
  • P. Kundu
    • 1
  1. 1.Computer Vision and Pattern Recognition UnitIndian Statistical InstituteCalcuttaIndia

Personalised recommendations