|Betreuer:||PhD Jan Koloda (Raum 01.178)|
|Hochschullehrer:||Prof. Dr.-Ing. André Kaup|
Many image reconstruction algorithms can be employed to reconstruct the lost pixels by dividing the image into a set of blocks. The unknown pixels within a block are then estimated using the available pixels within the same block. In turn, these already available pixels may have been previously obtained by estimation. In this situation, the order in which the blocks are processed will clearly condition the resulting reconstruction. The processing order (or filling order) for image reconstruction is crucial since reconstruction error can be propagated throughout the entire image.
As a reconstruction technique, the frequency selective extrapolation (FSE) algorithm will be considered. It is a powerful tool for image reconstruction and, by its nature, it is applied blockwise, so finding a suitable processing order is of utmost importance here.
In this thesis, the impact of the processing order on the resulting reconstruction is to be studied. Different criteria for establishing the processing order ought to be researched. The student shall implement the already existing methods as well as propose new ones, based, for instance, on the number of available pixels within a block or their spatial configuration. The performance shall be tested using a Matlab framework. The thesis is to be written in English.