A Fast Algorithm for Ill-posed Linear Inverse Problems

Master Thesis (2022)
Author(s)

F.F. Redel (TU Delft - Mechanical Engineering)

Contributor(s)

Peyman Mohajerinesfahani – Mentor (TU Delft - Team Peyman Mohajerin Esfahani)

M.R. Sheriff – Mentor (TU Delft - Team Peyman Mohajerin Esfahani)

Faculty
Mechanical Engineering
Copyright
© 2022 Floor Redel
More Info
expand_more
Publication Year
2022
Language
English
Copyright
© 2022 Floor Redel
Graduation Date
17-11-2022
Awarding Institution
Delft University of Technology
Programme
['Mechanical Engineering | Systems and Control']
Faculty
Mechanical Engineering
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

Ill-posed Linear Inverse Problems arise in various research domains, such as control engineering and image processing. Having a fast algorithm is a great benefit when working with high-dimensional signals, such as images. However, fast convergence and iterations with low computational complexity are challenging.
In this master thesis report, we propose an exact smooth reformulation of an ill-posed Linear Inverse Problem. Subsequently, we present a novel algorithm, the Fast Linear Inverse Problem Solver (FLIPS), associated with the new problem formulation. We show that in most metrics, the algorithm outperforms state-of-the-art methods like Chambolle-Pock (CP) and the Constrained Split Augmented Lagrangian Shrinkage Algorithm (C-SALSA) in terms of speed. Finally, associated with this algorithm, we present an open-source MATLAB package that includes the proposed algorithm and state-of-the-art methods.

Files

Master_thesis_report_Floor_Red... (pdf)
(pdf | 9.12 Mb)
- Embargo expired in 17-11-2023
License info not available