Point Spread Function Model of 3-Interaction Events for Super-MeV Compton Imaging Using Pixelated CdZnTe

Filtered backprojection presents a computationally cheap alternative to simple backprojection and maximum likelihood techniques to provide a high resolution image. It benefits from low memory requirements and the fact that it is a linear process which can easily be implemented in real time. The main component to filter design is the event response, or point spread function model. This work investigates the interaction mechanism in super-MeV Compton imaging to derive an analytical point spread function model. The model is specific for 3-interaction events and takes into account missequencing effects. It is then used in a spherical harmonic Wiener filter to reconstruct the images. When applying the filter with the analytical model, the image resolution improves from 33.6° average FWHM to 14.5° in a two source system. The analytical model's performance is poorer when compared to the 9.5° average FWHM of the Monte Carlo simulated model.