Browsing by Author "Djurabekova, Nargiza"

Sparse and limited angle tomography are common techniques used in computerized tomography to understand the inner working of live and inanimate objects with a reduced radiation dose. Unfortunately due to the lack of full data, these problems are often severely ill-posed and need powerful regularization strategies to create good reconstructions. One such tool is total variation, which creates good but patchy reconstructions. In this paper we analyze the use of non-local total variation as a way to improve the reconstruction quality. The analysis involves using both, simulated and real world data and comparing the TV reconstructions with the NLTV ones. The results leas us to believe that NLTV is particularly efficient and superior tool in sparse tomography, but further research is required to improve TV's limited-angle reconstructions.