On the choice of regularisation parameter in image restoration

  • J. W. Kay
Image Restoration And Enhancement
Part of the Lecture Notes in Computer Science book series (LNCS, volume 301)


This paper considers the application of the method of regularisation within the context of the restoration of degraded two-dimensional images. In particular, several recipes for choosing an appropriate degree of regularisation are described and their performance compared with reference to test-images. Some of these methods require the availability of a data-based noise-estimator; a neighbourhood noise estimator is proposed and its performance is discussed.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    T. Poggio, V. Torre and C. Koch (1985). Computational Vision and Regularisation Theory. Nature, Vol. 317, 314–319.PubMedGoogle Scholar
  2. [2]
    D.S. Sharman, K.A. Stewart and T.S. Durrani (1986). The use of Context in Image Restoration. Presented at the BPRA (Scotland) conference on Adaptive Image Analysis at Heriot-Watt University, May, 1986.Google Scholar
  3. [3]
    A.R. Davies and R.S. Anderssen (1986). Optimisation in the Regularisation of Ill-posed Problems. Journal of the Australian Mathematical Soceity (Series B), 28, 114–133.Google Scholar
  4. [4]
    D.M. Titterington (1985). Common Structure of Smoothing Techniques in Statistics. International Statistical Review, 53, 141–70.Google Scholar
  5. [5]
    P. Craven and G. Wahba (1979). Smoothing Noisy Data with spline functions: estimating the correct degree of smoothing by the method of generalised cross-validation. Number. Math., 31, 377–403.CrossRefGoogle Scholar
  6. [6]
    A.M. Thompson, J.C. Brown, J.W. Kay and D.M. Titterington (1987). A comparison of methods of choosing the smoothing parameter in image restoration by regularisation. In preparation.Google Scholar
  7. [7]
    C.W. Groetsch. The Theory of Tikhonov regularisation for Fredholm equations of the first kind. Research Notes in Mathematics No.105, Pitman.Google Scholar
  8. [8]
    P. Hall and D.M. Titterington (1987). Common structure of techniques for choosing smoothing parameters in regression problems. J.R. Statist. Soc. B., 49, No.2, 184–198.Google Scholar
  9. [9]
    P. Hall and D.M. Titterington (1986). On some Smoothing Techniques used in Image Restoration. J.R. Statist. Soc., 48, No. 3, 330–343.Google Scholar
  10. [10]
    V.F. Turchin (1967). Solution of the Fredholm equation of the first kind in a statistical ensemble of smooth functions. USSR Comp. Math. and Math. Phys., 7, 79–96.Google Scholar
  11. [11]
    V.F. Turchin (1968). Selection of an ensemble of smooth functions for the solution of the inverse problem. USSR Comput. Math. and Math. Phys., 8, 328–339.Google Scholar
  12. [12]
    G.H. Golub, M. Heath and G. Wahba (1979). Generalised cross-validation as a method for choosing a good ridge parameter. Technometrics, 21, 215–223. validation as a method for choosing a good ridge parameter. Technometrics, 21, 215–223.Google Scholar
  13. [13]
    A.K. Jain (1978). Fast Inversion of Banded Toeplitz Matrices by Circular Decompositions. IEEE Trans. A.S.S.P., Vol. ASSP-26, No. 2, 1978.Google Scholar
  14. [14]
    R.C. Gonzales & P. Wintz (1987). Digital Image Processing. Second Edition. Addison-Wesley.Google Scholar
  15. [15]
    R.M. Gray. On the Asymptotic Eigenvalue Distribution of Toeplitz Matrices. IEEE Trans. Inf. Th., Vol. IT-18, No. 6, Nov. 1972.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1988

Authors and Affiliations

  • J. W. Kay
    • 1
  1. 1.University of GlasgowScotland, UK

Personalised recommendations