Parameter optimization of a 3D SPECT parallel image reconstruction algorithm considering clinical aspects

OData support
Dr. Szlávecz Ákos
Department of Control Engineering and Information Technology

Parallel projection based SPECT (Singel Photon Emission Computed Tomography) is one of the most important functional imaging techniques nowadays. A gamma-emitting radionuclide is injected into the organ to be analysed. The distribution of the radionuclide is estimated through the spatial detection of the gamma-rays. On the basis of the distribution various diseases are diagnosed.

The measurement results are degraded by noise, artifacts are produced due to the non-homogeneous attenuation medium and the distance dependent spatial resolution of the imaging technique. The goal of the researches is to create high-quality reconstructed images based on the noisy data. Many analytical and iterative algorithms have been developed in order to reduce the impact of noise on the reconstructed image. In order to avoid the production of artifacts and improve the image quality we have to optimize the parameters of the 3D OSEM algorithm.

OSEM (Ordered Subset Expectation Maximization) is a new 3D iterative algorithm. Before the clinical application of the algorithm, need to carry out several tests either by phantoms or retrospective patient studies, in order to determinate the artifacts or imperfections of the new procedures.

The main goal of my work is to optimize the parameters of the 3D reconstruction algorithm for the investigated organs. especially for brain SPECT study have been selected for the parameter optimization procedure. Especially I should do it both for dedicated and multi-modality imaging systems. I studied the image quality with physical, anatomical phantoms and retrospective patient studies by changes of the parameters of the algorithm. The results has been tested by quantitative, semi-quantitative and qualitative way. The major consideration of the image quality were resolution, noise elimination and the running time of the algorithm for the optimal parametrization.

I present the highest quality reconstructed images and based on my studies I made a conclusion about the changes of the image quality by the two imaging systems.


Please sign in to download the files of this thesis.