Implementation of statistical 3D-PET image reconstruction with motion correction

摘要

The patient's movements during the measurement process induce an artifact in the reconstructed image in high resolution positron emission tomography (PET). An effective motion-correction method for statistical PET reconstruction has proposed. This method requires a considerable time to the three-dimensional global sensitivity image (GSI) which is a denominator of the ML-EM iteration formula. We proposed a practical approximation method for the calculation of the GSI by the rigid transformation images of the first motion-frame's GSI. From the result of the simulation, the proposed method reduced image artifacts as well as the computation time.