An alternative approach to the maximum-likelihood solution of deconvolution problems is presented. The resulting algorithms are faster converging than the conventional Richardson–Lucy and clean algorithms, as well as being more flexible when one is dealing with different types of noise. The performance of the algorithms on both Poisson and independent sensor noise is quantified.
© 1996 Optical Society of America
Original Manuscript: March 5, 1996
Manuscript Accepted: May 17, 1996
Published: October 1, 1996
R. G. Lane, "Methods for maximum-likelihood deconvolution," J. Opt. Soc. Am. A 13, 1992-1998 (1996)