Acceleration of maximum-likelihood image restoration for fluorescence microscopy and other noncoherent imagery

Holmes, Timothy; Liu, Yi-Hwa

doi:10.1364/josaa.8.000893

Cited by 65 publications

(49 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The EM algorithm (8) is identical to the Richardson-Lucy algorithm as derived by Richardson (1972). Although accelerated and regularized versions of the RichardsonLucy algorithm exist (Holmes & Liu, 1991;Joshi & Miller, 1993), none of these algorithms incorporate the background term bðxÞ.…”

Section: Restoration Methodsmentioning

confidence: 99%

“…In practice, this procedure is undesirable. Experiments (Holmes & Liu, 1991) show that the likelihood of a Richardson-Lucy estimate increases logarithmically as a function of the number of iterations. This growth makes the search for the maximum of the likelihood function extremely computationally expensive.…”

Section: Iterative Optimization: Where To Start and When To Stopmentioning

confidence: 99%

See 1 more Smart Citation

A quantitative comparison of image restoration methods for confocal microscopy

Kempen

Vliet

Verveer

et al. 1997

Journal of Microscopy

146

111

View full text Add to dashboard Cite

SummaryIn this paper, we compare the performance of three iterative methods for image restoration: the Richardson-Lucy algorithm, the iterative constrained Tikhonov-Miller algorithm (ICTM) and the Carrington algorithm. Both the ICTM and the Carrington algorithm are based on an additive Gaussian noise model, but differ in the way they incorporate the non-negativity constraint. Under low light-level conditions, this additive (Gaussian) noise model is a poor description of the actual photon-limited image recording, compared with that of a Poisson process. The RichardsonLucy algorithm is a maximum likelihood estimator for the intensity of a Poisson process. We have studied various methods for determining the regularization parameter of the ICTM and the Carrington algorithm and propose a (Gaussian) prefiltering to reduce the noise sensitivity of the Richardson-Lucy algorithm. The results of these algorithms are compared on spheres convolved with a point spread function and distorted by Poisson noise. Our simulations show that the Richardson-Lucy algorithm, with Gaussian prefiltering, produces the best result in most of the tests. Finally, we show an example of how restoration methods can improve quantitative analysis: the total amount of fluorescence inside a closed object is measured in the vicinity of another object before and after restoration.

show abstract

Section: Restoration Methodsmentioning

confidence: 99%

Section: Iterative Optimization: Where To Start and When To Stopmentioning

confidence: 99%

A quantitative comparison of image restoration methods for confocal microscopy

Kempen

Vliet

Verveer

et al. 1997

Journal of Microscopy

146

111

View full text Add to dashboard Cite

show abstract

“…Performing RL as described in (8) often results in a very slow convergence. For the acceleration of RL, several methods have been suggested [7,[19][20][21]. Among them, the technique in [7] has been noted for its success.…”

Section: Richardson-lucy Iterationmentioning

confidence: 99%

Block-iterative Richardson-Lucy methods for image deblurring

Lee

2015

J Image Video Proc.

View full text Add to dashboard Cite

In this paper, we extend the Richardson-Lucy (RL) method to block-iterative versions, separated BI-RL, and interlaced BI-RL, for image deblurring applications. We propose combining algorithms for separated BI-RL to form block artifact-free output images from separately deblurred block images. For interlaced BI-RL to accelerate the iteration, we propose an interlaced block-iteration algorithm on down-sampled blocks of the observed image. Simulation studies show that separated BI-RL and interlaced BI-RL achieve desired goals in Gaussian and diagonal deblurrings.

show abstract

“…However, although the RL algorithm is very popular, it has drawbacks in terms of the restoration quality and implementation time for 3D images. To address the slow convergence of the RL algorithm many acceleration methods have been proposed, which try to select step size k  [5], or adjust the search direction   k Cx  [6,7].…”

Section: B Gradient Projection and Related Workmentioning

confidence: 99%