Region of Interest Image Reconstruction using IR Tools Open Access

Huang, Xiaoyi (Spring 2019)

Permanent URL:


In this thesis, we consider large-scale, ill-posed inverse problems that arise in image processing applications. These problems can be modeled as linear systems, where the matrix that models the forward problem is extremely ill-conditioned. In addition, the observed data is contaminated by noise. Due to the ill-conditioning of the matrix and the presence of noise in the observed data, it is necessary to employ regularization to compute a meaningful approximation of the solution.

Although there has been some very effective but expensive image reconstruction algorithm, those algorithms cannot be applied to large images because of their cost. This thesis focuses on using inexpensive, fast methods to obtain an initial image reconstruction first, combined with more expensive methods to improve specific subimages, called region of interest (ROI) areas in the image.

Table of Contents

1. Introduction ---------------------------------------------------------------- 1

1.1 Mathematical Background -------------------------------------------------2

1.2 Conditioning of the Problem ---------------------------------------------- 4

1.3 Computational Difficulties ------------------------------------------------ 6

2. Regularization --------------------------------------------------------------9

2.1 Truncated SVD -----------------------------------------------------------10

2.2 Tikhonov Regularization -------------------------------------------------10

3. Image Deblurring ----------------------------------------------------------13

3.1 Spacially Invariant Blur --------------------------------------------------14

3.2 Spacially Variant Blur ----------------------------------------------------20

4. Iterative Methods for Image Deblurring -----------------------------------22

4.1 Iterative Solvers in IR Tools ----------------------------------------------23

4.2 Solve 2D Image Deblurring Problems ------------------------------------ 25

5. Region of Interest Computations ------------------------------------------35

5.1 Extract ROI subimages ---------------------------------------------------36

5.2 ROI Experiments ---------------------------------------------------------45

6. Concluding Remarks ------------------------------------------------------ 57

About this Honors Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
  • English
Research Field
Committee Chair / Thesis Advisor
Committee Members
Last modified

Primary PDF

Supplemental Files