Rank-deficient and Discrete Ill-posed Problems
Number of credits: 9 hp
Examiner: Tommy Elfving
Organisation: Project 1. Implementation of the algorithms CGLS and CGMN in the SNARK programming package. Tests and comparisons of these algorithms with Landsweber and Art iteration on problems from image reconstruction and electron microscope image reconstruction. Behavior using noisy and noise free data. Stopping criteria.
Project 2. Comparison using CGLS resp CG on regularized problem. Choice of regularization parameter. Effect of using regularizing preconditioners.
Examination: Reading course. The student will present the contents of the literature at 8 bi-weekly seminars. Student will also do projects/assignments related to the course.
Last updated: 2014-04-29