Abstract:SIAM Journal on Imaging Sciences, Volume 17, Issue 2, Page 1206-1254, June 2024. This paper presents a novel stochastic optimization methodology to perform empirical Bayesian inference in semi-blind image deconvolution problems. Given a blurred image and a parametric class of possible operators, the proposed optimization approach automatically calibrates the parameters of the blur model by maximum marginal likelihood estimation, followed by (non-blind) image deconvolution by maximum a posteriori estimation conditionally to the estimated model parameters. In addition to the blur model, the proposed approach also automatically calibrates the noise level as well as any regularization parameters. The marginal likelihood of the blur, noise, and regularization parameters is generally computationally intractable, as it requires calculating several integrals over the entire solution space. Our approach addresses this difficulty by using a stochastic approximation proximal gradient optimization scheme, which iteratively solves such integrals by using a Moreau–Yosida regularized unadjusted Langevin Markov chain Monte Carlo algorithm. This optimization strategy can be easily and efficiently applied to any model that is log-concave and by using the same gradient and proximal operators that are required to compute the maximum a posteriori solution by convex optimization. We provide convergence guarantees for the proposed optimization scheme under realistic and easily verifiable conditions and subsequently demonstrate the effectiveness of the approach with a series of deconvolution experiments and comparisons with alternative strategies from the state of the art

Lp (1≤p≤∞) BLIND ADAPTIVE DECONVOLUTION AND ITS APPLICATIONS

L∞ Blind Deconvolution for the Generalized AR Model.

Multi-frame Blind Deconvolution Using Sparse Priors

Blind Deconvolution Using an Improved L0 Sparse Representation

Identifiability of Multichannel Blind Deconvolution and Nonconvex Regularization Algorithm

A Blind Deconvolution Method for Attenuative Materials Based on Asymmetrical Gaussian Model.

Non-blind image deconvolution using natural image gradient prior

High quality non-blind image deconvolution using the Fields of Experts prior

Blind Image Deblurring Using Spectral Properties of Convolution Operators

Rapid, Robust, and Reliable Blind Deconvolution Via Nonconvex Optimization

A L-0 Sparse Analysis Prior for Blind Poissonian Image Deconvolution

Blind Deconvolution using Convex Programming

Generalized variational framework with minimax optimization for parametric blind deconvolution

Fundamental Limits of Blind Deconvolution Part I: Ambiguity Kernel

Nonlinear Blind Deconvolution Based on Generalized Normalized l p / l q Norm for Early Fault Detection

Blind Image Deblurring by Spectral Properties of Convolution Operators

Sparse Blind Deconvolution with Nonconvex Optimization for Ultrasonic NDT Application

Marginal Likelihood Estimation in Semiblind Image Deconvolution: A Stochastic Approximation Approach

Blind Image Deconvolution Assisted By Robust Non-Gaussian Priors

Adaptive Minimum Variance Methods for Direct Blind Multichannel Equalization

Gradient-based discriminative modeling for blind image deblurring