Scatter correction for positron volume imaging using analytic simulation

The removal of interslice septa in positron volume imaging results in an increase in the fraction of Compton-scattered gamma rays detected, requiring the use of more accurate correction methods to compensate adequately for the scattered rays. Three scatter correction methods for volume imaging which use the source distribution to generate scatter estimates were implemented. Two of the methods use convolution in the 2-D projections to estimate scatter; one uses a spatially varying scatter function, the other a spatially invariant scatter function. The third uses analytic simulation to estimate the scatter in the 2-D projections from a reconstructed 3-D image. The methods are applied to Monte Carlo simulated data to study the source distribution dependence of scatter correction. Scatter correction improves when the variation of the scatter function with source position is taken into account. The image-to-projection transformation method produces significantly better scatter estimates for small objects than either the spatially varying or the spatially invariant convolution methods.<>