Total variation-penalized Poisson likelihood estimation for ill-posed problems

Article

Abstract

The noise contained in data measured by imaging instruments is often primarily of Poisson type. This motivates, in many cases, the use of the Poisson negative-log likelihood function in place of the ubiquitous least squares data fidelity when solving image deblurring problems. We assume that the underlying blurring operator is compact, so that, as in the least squares case, the resulting minimization problem is ill-posed and must be regularized. In this paper, we focus on total variation regularization and show that the problem of computing the minimizer of the resulting total variation-penalized Poisson likelihood functional is well-posed. We then prove that, as the errors in the data and in the blurring operator tend to zero, the resulting minimizers converge to the minimizer of the exact likelihood function. Finally, the practical effectiveness of the approach is demonstrated on synthetically generated data, and a nonnegatively constrained, projected quasi-Newton method is introduced.

Keywords

Total variation regularization Ill-posed problems Maximum likelihood estimation Image deblurring Nonnegatively constrained minimization 

PACS

02.30.Zz 02.50.-r 07.05.Pj 

References

  1. 1.
    Acar, R., Vogel, C.R.: Analysis of bounded variation penalty methods for ill-posed problems. Inverse Problems 10, 1217–1229 (1994)MATHCrossRefMathSciNetGoogle Scholar
  2. 2.
    Asaki, T., Chartrand, R., Le, T.: A Variational Approach to Reconstructing Images Corrupted by Poisson Noise. UCLA CAM Report 05-49 (November 2005)Google Scholar
  3. 3.
    Bardsley, J.M.: A limited memory, quasi-newton preconditioner for nonnegatively constrained image reconstruction. J. Opt. Soc. Amer. A 21(5), 724–731 (2004)CrossRefGoogle Scholar
  4. 4.
    Bardsley, J.M., Laobeul, N.: Tikhonov regularized Poisson likelihood estimation: theoretical justification and a computational method. Inverse Probl. Sci. Eng. 16(2), 199–215 (January 2008)CrossRefMathSciNetGoogle Scholar
  5. 5.
    Bardsley, J.M., Vogel, C.R.: A nonnnegatively constrained convex programming method for image reconstruction. SIAM J. Sci. Comput. 25(4), 1326–1343 (2004)CrossRefMathSciNetGoogle Scholar
  6. 6.
    Berterro, M., Boccacci, P.: Introduction to Inverse Problems in Imaging. Institute of Physics Publishing, Bristol (1998)CrossRefGoogle Scholar
  7. 7.
    Bertsekas, D.P.: Projected newton methods for optimization problems with simple constraints. SIAM J. Control Optim. 20, 221–246 (1982)MATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Chambolle, A., Lions, P.-L.: Image recovery via total variation minimization and related problems. Numer. Math. 76(2), 167–188 (1997)MATHCrossRefMathSciNetGoogle Scholar
  9. 9.
    Conway, J.B.: A Course in Functional Analysis, 2nd edn. Springer, Heidelberg (1990)MATHGoogle Scholar
  10. 10.
    Evans, L.C., Gariepy, R.F.: Measure Theory and Fine Properties of Functions. Studies in Advanced Mathematics. CRC, Boca Raton (1992)Google Scholar
  11. 11.
    Giusti, E.: Minimal Surfaces and Functions of Bounded Variation. Monographs in Mathematics, vol. 80. Birkhauser Verlag, Basel (1984)Google Scholar
  12. 12.
    Green, M.: Statistics of images, the TV algorithm of Rudin-Osher-Fatemi for image denoising, and an improved denoising algorithm. CAM Report 02-55, UCLA (October 2002)Google Scholar
  13. 13.
    Kelley, C.T.: Iterative Methods for Optimization. SIAM, Philadelphia (1999)MATHGoogle Scholar
  14. 14.
    Le, T., Chartrand, R., Asaki, T.J.: A variational approach to reconstructing images corruted by poisson noise. J. Math. Imaging Vision 27(3), 257–263 (2007)CrossRefMathSciNetGoogle Scholar
  15. 15.
    Moré, J.J., Toraldo, G.: On the solution of large quadratic programming problems with bound constraints. SIAM J. Optim. 1, 93–113 (1991)MATHCrossRefMathSciNetGoogle Scholar
  16. 16.
    Nocedal, J., Wright, S.: Numerical Optimization. Springer, Heidelberg (1999)MATHCrossRefGoogle Scholar
  17. 17.
    Nowak, R., Kolaczyk, E.: A statistical multiscale framework for poisson inverse problems. IEEE Trans. Inform. Theory 46(5), 1811–1825 (2000)MATHCrossRefMathSciNetGoogle Scholar
  18. 18.
    Rudin, L.I., Osher, S., Fatemi, E.: Nonlinear total variation based noise removal algorithms. Phys. D 60(1–4), 259–268 (1992)MATHCrossRefGoogle Scholar
  19. 19.
    Sardy, S., Antoniadis, A., Tseng, P.: Automatic smoothing with wavelets for a wide class of distributions. J. Comput. Graph. Statist. 13(2), 1–23 (2004)MathSciNetGoogle Scholar
  20. 20.
    Sardy, S., Tseng, P.: On the statistical analysis of smoothing by maximizing dirty Markov random field posterior distributions. J. Amer. Statist. Assoc. 99(465), 191–204(14) (2004)MATHCrossRefMathSciNetGoogle Scholar
  21. 21.
    Snyder, D.L., Hammoud, A.M., White, R.L.: Image recovery from data acquired with a charge-coupled-device camera. J. Opt. Soc. Amer. A 10, 1014–1023 (1993)CrossRefGoogle Scholar
  22. 22.
    Snyder, D.L., Helstrom, C.W., Lanterman, A.D., Faisal, M., White, R.L.: Compensation for readout noise in CCD images. J. Opt. Soc. Amer. A 12, 272–283 (1995)CrossRefGoogle Scholar
  23. 23.
    Tikhonov, A.N., Goncharsky, A.V., Stepanov, V.V., Yagola, A.G.: Numerical Methods for the Solution of Ill-Posed Problems. Kluwer Academic, Boston (1990)MATHGoogle Scholar
  24. 24.
    Yu, D.F., Fessler, J.A.: Edge-preserving tomographic reconstruction with nonlocal regularization. IEEE Trans. Med. Imag. 21(2), 159–173 (2002)CrossRefGoogle Scholar
  25. 25.
    Vogel, C.R.: Computational Methods for Inverse Problems. SIAM, Philadelphia (2002)MATHGoogle Scholar
  26. 26.
    Vogel, C.R., Oman, M.E.: A fast, robust algorithm for total variation based reconstruction of noisy, blurred images. IEEE Trans. Image Process. 7, 813–824 (1998)MATHCrossRefMathSciNetGoogle Scholar
  27. 27.
    Zeidler, E.: Applied Functional Analysis: Main Principles and their Applications. Springer, New York (1995)MATHGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC. 2008

Authors and Affiliations

  1. 1.Department of Mathematical SciencesUniversity of MontanaMissoulaUSA
  2. 2.Division of Mathematics and Computer ScienceClarkson UniversityPotsdamUSA

Personalised recommendations