CORRIGENDUM: The expectation-maximization algorithm for ill-posed integral equations: a convergence analysis

The expectation-maximization (EM) algorithm is a convenient tool for approximating maximum likelihood estimators in situations when available data are incomplete, as is the case for many inverse problems. Our focus here is on the continuous version of the EM algorithm for a Poisson model, which is known to perform unstably when applied to ill-posed integral equations. We interpret and analyse the EM algorithm as a regularization procedure. We show weak convergence of the iterates to a solution of the equation when exact data are considered. In the case of perturbed data, similar results are established by employing a stopping rule of discrepancy type under boundedness assumptions on the problem data.

[1]  R. A. Leibler,et al.  On Information and Sufficiency , 1951 .

[2]  V. Morozov.  On the solution of functional equations by the method of regularization , 1966 .

[3]  M. Perlman.  Jensen's inequality for a convex vector-valued function on an infinite-dimensional space , 1974 .

[4]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[5]  A. Kondor,et al.  Method of convergent weights — An iterative procedure for solving Fredholm's integral equations of the first kind , 1983 .

[6]  L. Shepp,et al.  Maximum Likelihood Reconstruction for Emission Tomography , 1983, IEEE Transactions on Medical Imaging.

[7]  K. Lange,et al.  EM reconstruction algorithms for emission and transmission tomography. , 1984, Journal of computer assisted tomography.

[8]  L. Shepp,et al.  A Statistical Model for Positron Emission Tomography , 1985 .

[9]  R. S. Anderssen,et al.  The Linear Functional Strategy for Improperly Posed Problems , 1986 .

[10]  B. Schorr,et al.  On an iterative method for a class of integral equations of the first kind , 1987 .

[11]  E. Veklerov,et al.  Stopping Rule for the MLE Algorithm Based on Statistical Hypothesis Testing , 1987, IEEE Transactions on Medical Imaging.

[12]  B. Schorr,et al.  On properties of the iterative maximum likelihood reconstruction method , 1989 .

[13]  J. D. Wilson,et al.  A smoothed EM approach to indirect estimation problems, with particular reference to stereology and emission tomography , 1990 .

[14]  Peter Maass,et al.  A mollifier method for linear operator equations of the first kind , 1990 .

[15]  Jonathan M. Borwein,et al.  Convergence of Best Entropy Estimates , 1991, SIAM J. Optim..

[16]  Alfredo N. Iusem.  Convergence analysis for a multiplicatively relaxed EM algorithm , 1991 .

[17]  H. N. Mülthei.  Iterative continuous maximum‐likelihood reconstruction method , 1992 .

[18]  Y. Vardi,et al.  From image deblurring to optimal investments : maximum likelihood solutions for positive linear inverse problems , 1993 .

[19]  Robert S. Anderssen,et al.  On the stabilization inherent in the EMS algorithm , 1994 .

[20]  P. Eggermont,et al.  Maximum penalized likelihood estimation and smoothed em algorithms for positive integral equatioins of the first kind , 1996 .

[21]  P. P. B. Eggermont,et al.  Nonlinear Smoothing and the EM Algorithm for Positive Integral Equations of the First Kind , 1999 .

[22]  E. Resmerita.  Regularization of ill-posed problems in Banach spaces: convergence rates , 2005 .

[23]  A. Munk,et al.  A multi-scale stopping criterion for MLEM reconstructions in PET , 2006, 2006 IEEE Nuclear Science Symposium Conference Record.

[24]  R. Anderssen,et al.  Joint additive Kullback–Leibler residual minimization and regularization for linear inverse problems , 2007 .