Iterative regularization algorithms for constrained image deblurring on graphics processors

Ruggiero, Valeria; Serafini, Thomas; Zanella, Riccardo; Zanni, Luca

doi:10.1007/s10898-009-9516-x

Cited by 13 publications

(10 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We remark that in these examples the same results are obtained by means of the discrepancy function (15). Since SGP has been already implemented on graphics processors (GPU) [29], implementing the discrepancy stopping rule might provide a very fast tool for image deblurring with automatic stopping of iterations. Potential applications are open to image deblurring in astronomy and microscopy.…”

Section: Deblurringsupporting

confidence: 60%

A discrepancy principle for Poisson data

et al. 2010

View full text Add to dashboard Cite

In applications of imaging science, such as emission tomography, fluorescence microscopy and optical/infrared astronomy, image intensity is measured via the counting of incident particles (photons, γ-rays, etc). Fluctuations in the emission-counting process can be described by modeling the data as realizations of Poisson random variables (Poisson data). A maximum-likelihood approach for image reconstruction from Poisson data was proposed in the mid-1980s. Since the consequent maximization problem is, in general, ill-conditioned, various kinds of regularizations were introduced in the framework of the so-called Bayesian paradigm. A modification of the well-known Tikhonov regularization strategy results in the data-fidelity function being a generalized Kullback-Leibler divergence. Then a relevant issue is to find rules for selecting a proper value of the regularization parameter. In this paper we propose a criterion, nicknamed discrepancy principle for Poisson data, that applies to both denoising and deblurring problems and fits quite naturally the statistical properties of the data. The main purpose of the paper is to establish conditions, on the data and the imaging matrix, ensuring that the proposed criterion does actually provide a unique value of the regularization parameter for various classes of regularization functions. A few numerical experiments are performed to demonstrate its effectiveness. More extensive numerical analysis and comparison with other proposed criteria will be the object of future work.

show abstract

Section: Deblurringsupporting

confidence: 60%

A discrepancy principle for Poisson data

et al. 2010

View full text Add to dashboard Cite

show abstract

“…The GPU implementation of the algorithms is obtained by performing on the parallel device the main computational tasks (FFT computations and vector-vector operations) of each iteration. In this way, by running the same number of iterations with the serial and the parallel versions of the deconvolution algorithms, essentially the same reconstruction accuracy is obtained [12,29,39]. For each image size, the results reported in Table 4 and 5 are obtained as average values of ten test problems corresponding to different noise realizations.…”

Section: Gpu Testsmentioning

confidence: 83%

Limited-memory scaled gradient projection methods for real-time image deconvolution in microscopy

Porta

Zanella

Zanghirati

et al. 2015

Communications in Nonlinear Science and Numerical Simulation

View full text Add to dashboard Cite

“…Thus the computational complexity depends on the number of the matrix-vector products Hx E ; H T z. These features make SGP well-suited to be implemented also on graphic processing units (GPUs) or on multiprocessor systems [32].…”

Section: Statement Of the Problemmentioning

confidence: 99%

Reprint of Inexact Bregman iteration for deconvolution of superimposed extended and point sources

Benfenati

Ruggiero

2015

Communications in Nonlinear Science and Numerical Simulation

View full text Add to dashboard Cite

Iterative regularization algorithms for constrained image deblurring on graphics processors

Abstract: Image deblurring, Scaled gradient projection method, Graphics processing units,

Cited by 13 publications

References 26 publications

A discrepancy principle for Poisson data

A discrepancy principle for Poisson data

Limited-memory scaled gradient projection methods for real-time image deconvolution in microscopy

Reprint of Inexact Bregman iteration for deconvolution of superimposed extended and point sources

Contact Info

Product

Resources

About