This paper describes implementation of controlled random search procedure employed to optimize and evaluate Bayesian image reconstructions algorithms with bandlimited basis functions. Median root prior (MRP) and its generalizations, i.e. the L-filter and hybrid MRP were optimized and compared to the Huber penalty. The coefficient controlling amount/weight of the penalty and parameter controlling width ofthe bandlimited basis functions were optimized for different amounts of noise in the data. The reconstruction methods were accelerated using ordered subsets principle. Taking into account quantitative accuracy of the reconstructions, basic MRP with bandlimited basis functions is a practical altemative to the L-filter or hybrid MRP.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.