Scribd
Upload
32 views2 pages

- δ) : Gaussian Markov random fields - . - . - . - . - 54 - δ) : Laplace Markov random fields - . - . - . - . - . 65

This document is the contents page for a book on inverse problems. It lists the six chapters of the book and their section headings. Chapter 1 introduces characteristics of inverse problems and provides an illustrative example. Chapter 2 describes regularization by spectral filtering methods and parameter selection. Chapter 3 presents two-dimensional test cases of image deblurring and computed tomography. Chapter 4 discusses Bayes' law, Markov random field priors, and maximum a posteriori estimation. Chapter 5 covers Markov chain Monte Carlo methods for linear inverse problems. Chapter 6 extends these methods to nonlinear inverse problems.

Uploaded by

Nguyễn Quang Huy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
32 views2 pages

- δ) : Gaussian Markov random fields - . - . - . - . - 54 - δ) : Laplace Markov random fields - . - . - . - . - . 65

This document is the contents page for a book on inverse problems. It lists the six chapters of the book and their section headings. Chapter 1 introduces characteristics of inverse problems and provides an illustrative example. Chapter 2 describes regularization by spectral filtering methods and parameter selection. Chapter 3 presents two-dimensional test cases of image deblurring and computed tomography. Chapter 4 discusses Bayes' law, Markov random field priors, and maximum a posteriori estimation. Chapter 5 covers Markov chain Monte Carlo methods for linear inverse problems. Chapter 6 extends these methods to nonlinear inverse problems.

Uploaded by

Nguyễn Quang Huy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Contents

Preface vii

1 Characteristics of Inverse Problems 1


1.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The least squares estimator . . . . . . . . . . . . . . . . . . . . . 6
1.3 The statistical properties of xLS and ill-posedness . . . . . . . . . 8
1.4 An illustrative example . . . . . . . . . . . . . . . . . . . . . . . 11
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2 Regularization by Spectral Filtering 15


2.1 Spectral filtering methods . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Regularization parameter selection methods . . . . . . . . . . . . 19
2.3 Periodic and data-driven boundary conditions . . . . . . . . . . . 25
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3 Two-Dimensional Test Cases 33


3.1 Two-dimensional image deblurring . . . . . . . . . . . . . . . . . 33
3.2 Computed tomography . . . . . . . . . . . . . . . . . . . . . . . 42
3.3 The preconditioned conjugate gradient iteration . . . . . . . . . . 45
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4 Bayes’ Law, Markov Random Field Priors, and MAP Estimation 53


4.1 Bayes’ law and regularization . . . . . . . . . . . . . . . . . . . . 53
4.2 Choosing p(x|δ): Gaussian Markov random fields . . . . . . . . . 54
4.3 Choosing p(x|δ): Laplace Markov random fields . . . . . . . . . . 65
4.4 The infinite-dimensional limit . . . . . . . . . . . . . . . . . . . . 68
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

5 Markov Chain Monte Carlo Methods for Linear Inverse Problems 77


5.1 Sampling from high-dimensional Gaussian random vectors . . . . 77
5.2 Hierarchical modeling of λ and δ and sampling from p(x, λ, δ|b) . 81
5.3 Alternative MCMC methods for sampling from p(x, λ, δ|b) . . . . 89
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

6 Markov Chain Monte Carlo Methods for Nonlinear Inverse Problems 105
6.1 A general setup for nonlinear inverse problems . . . . . . . . . . . 105
6.2 Levenburg–Marquardt nonlinear least squares optimization . . . . 106

v
vi Contents

6.3 Randomize-then-optimize as a proposal for Metropolis–Hastings . 109


6.4 Nonlinear test cases . . . . . . . . . . . . . . . . . . . . . . . . . 112
6.5 Hierarchical modeling of λ and δ and sampling from p(x, λ, δ|b) . 118
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

Bibliography 125

Index 131

You might also like