The job of doctors have become much easier thanks to the development of medical imaging systems. These systems use a variety of technologies, and each one of them is designed for a different purpose. The PET (Positron Emission Tomography) is one of such imaging systems, one of the newer ones, still under development.
One of the directions of the development is enabling dynamic reconstruction. With its help, doctors can examine the behaviour of an organ or a substance, while it is inside a living organism. This simplifies both the research and the diagnosis.
Our goal was to implement the dynamic reconstruction. In order to achieve our goal, we altered the storage of the measurement data and the reconstruction algorithm. To handle the temporal property we tried different representational models, examined the possibility of algorithmic level optimalization and its effect on the quality of the reconstruction.
Finally we summarized the results and discussed which approach meets our expectations.