FWF-Project P19496
Inverse problems with sparsity constraints
| Project Leader: | PD Dr. Ronny Ramlau |
| Research scientist: | DI. Stephan Anzengruber |
Project Abstract
The combination of sparse signal reconstruction and ill-posed inverse problems is a growing area of research and will be investigated in this project. More specifically we are dealing with the development of new regularization techniques for nonlinear ill-posed inverse problems with sparsity constraints. The development of algorithms for the case of linear problems in Tikhonov functional form has been a main focus of research throughout the past years. A solution for the linear case was presented four years ago by Daubechies et al. [DDDM04]. In 2005 we presented such a concept for nonlinear problems and first results look very promising even though we are far from having a complete analysis of this case. Therefore the first goal of this project will be to merge together the topics of sparse signal reconstruction and nonlinear inverse problems and to create thereby a foundation for a number of applications, such as the analysis of the dynamics of cellular networks, medical imaging, and rotor dynamics. All these three applications result in nonlinear ill-posed problems whose solutions are assumed to be sparse. Therefore the methods to be developed in this project will contribute to removing the flaws of the algorithms currently in use. First simple results show that the natural approach will be to aim for iterative algorithms. With this in mind the canonical work program for this project is:- Development of algorithms
- Convergence properties (weak, strong, conditions which guarantee convergence)
- Acceleration of the methods
- Analysis of the convergence rates
- Establishment of the necessary regularization theory (parameter choice rules)
- Implementation of the algorithms (in cooperation with industrial project partners) and evaluation.
Keywords and AMS Classification
- inverse problems
- sparsity
- nonlinear ill-posed problems
- parameter choice rules
The Institute is named after the famous Austrian mathematician Johann Radon (1887-1956)
Medieninhaber:
Österreichische Akademie der Wissenschaften
Juristische Person öffentlichen Rechts (BGBl 569/1921 idF BGBl I 130/2003)
Dr. Ignaz Seipel-Platz 2, 1010 Wien
Diese Website dient zur Information über die wissenschaftlichen Aktivitäten der Österreichischen Akademie der Wissenschaften und setzt somit den gesetzlichen Auftrag um, die Wissenschaft in jeder Hinsicht zu fördern.
This RICAM page was made with 100% valid HTML & CSS - Send comments to Webmaster
Today's date and time is 03/18/10 - 18:03 CET and this file ( /projects/sparsity/index.html ) was last modified on 04/09/08 - 17:12 CEST