Modeling Scatter in Multiple Energy Windows to Incorporate Scattered Radiation Information in Single Photon Emission Computed Tomography Imaging

Abstract

Using scattered radiation as useful information to improve radioisotope image quality is a topic attracting many researchers. Some reports showed that incorporating scattered components offers a possibility to improve image quality. The general method is modeling scatter in multiple energy windows and incorporating that information into the reconstruction process. However, what the performance will be and how noise will behave when using scattered radiation in reconstruction are not yet well answered. In this paper, we investigate a method of modeling scatter in multiple energy windows in cases of a few projection views. The system performance is analyzed using singular value decomposition and resolution kernels. For noise behavior investigation, reconstructions are accomplished by estimating the variance of reconstructed voxel values and the effectiveness of using scatter is evaluated by resolution kernel analysis. The results show there are improvements in normalized mean-square error of the images and the resolution kernels. When photon counts fall below about one million, it is still effective to use scatter for some cases of a few projections.