Direct 3D image reconstruction for the energy subtraction Compton scattering camera (ESCSC) with spatially invariant point spread function

A direct 3D image reconstruction technique has been implemented for the energy subtraction Compton scattering camera (ESCSC) data. By simulating a point source with MCNP and projecting cone surfaces that result from preferred events (defined as photons that undergo a single Compton scattering in the primary detector and then a photoelectric absorption in the secondary detector) into the source space, a point spread function (PSF) describing the characteristics of the imaging system was determined. Based on the Fourier convolution theorem, the PSF was considered as a system response function h(r), and an image was reconstructed via deconvolution using the Fourier transform. For the ESCSC, different weighting schemes can result in spatially invariant or spatially variant PSFs. In this paper, a weighting scheme leading to a nearly spatially invariant PSF was chosen for the backprojection operation. Using MCNP simulated data, several distributed sources were used to test this direct technique. The images of the distributed sources were reconstructed without the existence of simulated noise. Results show that the reconstructed images give reasonable dimensional information of objects, however the intensity recovery is not uniform through the entire object space