SC²S Colloquium - September 23, 2015
|Date:||September 23, 2015|
|Time:||3:00 pm, s.t.|
Steen Müller: Entwicklung und Implementierung eines Fahrsimulators
Annamaria Petreus: A Comparison of Several Tikhonov Regularization Methods in the Image Deblurring Process
Discrete ill-posed problems represent a challenge, because their solution is very sensitive to both errors in the data and round-off errors that occur during the solution process. The Tikhonov regularization method replaces the given discrete ill-posed problem by a penalized least-squares problem and seeks to yield a much stable solution. An important role in the solution process play the choice of the regularization matrix and the potential filter factors introduced in the penalty term.
One of the fields of application of Tikhonov regularization is the image deblurring process. This thesis is concerned with the singular value decomposition (SVD) of the blur operator and the Tikhonov regularization method in the vector space. Several regularization matrices and filter factors have been considered. Their influence on the final solution has been examined with MATLAB programs.