Implementation and verification of the Convolution Subtraction scatter correction algorithm applied for emission tomography imaging

OData support
Dr. Benyó Balázs István
Department of Control Engineering and Information Technology

Nowadays nuclear imaging devices play an important role in the modern healthcare, e.g. oncology, radiology or neurology. A substantial task for high quality diagnosis is to assure the high accuracy of nuclear medical images.

During examination a radiopharmacon is injected in the patient that is accumulating in specific location in the body (e.g. the tissue of cancer). The results of the radioactive decay are directly (SPECT) or indirectly (PET) high energy photons, called gamma photons. These gamma photons are detected, and the spatial distribution of the isotope can be reconstructed. One of the most important physical effects resulting in distortions in reconstructed images is the Compton scattering.

The aim of this work is to implement a scatter correction algorithm and the appropriate calibration method. The Convolution Subtraction scatter correction method is an iterative process that eliminates the scattered photons utilizing deconvolution technique.

The method was originally designed for SPECT devices, but with a slight modification it works for PET as well.

The dependency of the scatter on the radial distance from rotation axis is researched, as well. The paper contains the description and analysis of a scatter correction method while considering the radial distance dependency of the scatter model.

The method has been tested with simulation studies created with the GATE Monte Carlo simulation tool. Current results show an improved image quality.


Please sign in to download the files of this thesis.