Recently the Robotics Laboratory of Near East University joined research on medical imaging – namely, on new generation PET (positron emission tomography) scanners within GATE (Geant4 Application for Tomographic Emission) collaboration.
The goal of the research is to attain spatial resolution about 1 mm, whereas the state-of-the-art devices provide only several (4-5) mm. Our activity has come out with a new Matlab-implemented algorithm for the reconstruction of PET image. The algorithm, based on the so-called direct Fourier approach, provides fully 3-dimensional imaging: it means that all raw data are processed, and no simplifying techniques are used (such techniques, which we can also think of as “tricks”, simplify implementation and reduce computation time, but all at the cost of image quality). Raw data are generated by the GATE PET simulator – a software package, allowing to design PET scanner and estimate its performance without manufacturing the real device.
Below are the results of image reconstruction with NEMA phantom employed as the positron-emitting body. NEMA phantom is some standard object used to evaluate image quality.