ARTIFICIAL NEURAL-NETWORK AS A TOOL TO COMPENSATE FOR SCATTER AND ATTENUATION IN RADIONUCLIDE IMAGING

This study investigates the ability of artificial neural networks (ANN ) to simultaneously correct for attenuation and Compton scattering in scintigraphic imaging. Methods: Three sets of experiments are conducte d using images of radioactive sources with various shapes and distribu tions in a homogeneous medium. Numerical Monte Carlo simulations and p hysical phantom acquisitions of radioactive geometric sources provide the basic material for correction. Our method is based on the followin g assumptions: information needed to correct for scattering can be ext racted from the energy spectrum at each pixel without any assumption c oncerning the source distribution, and two diametrically opposed energ y spectrum acquisitions yield enough information on the source locatio n in the diffusing medium for simultaneous correction for attenuation and scattering. Results: Qualitative and quantitative evaluations of s catter correction by ANN demonstrate its ability to perform scatter co rrection from the energy spectra observed in each pixel, By using the energy spectra of incident photons detected in two diametrically oppos ed images, multilayer neural networks are able to perform a proper res titution of projection images without any assumption on geometry or po sition of radioactive sources in simple geometric cases. ANN correctio ns compare favorably to those provided by five of the most popular met hods. A satisfying correction of both scatter and attenuation is obser ved for a human pelvis scan obtained during routine clinical practice. Conclusion: An ANN is an efficient tool for attenuation and Compton s cattering in simple model cases, The results obtained for routine scin tigrams in a much more complex situation are strong incentives for per forming further studies.