Abstract
An energy-based scatter correction for SPECT and planar images is quantitatively evaluated using measured Rollo phantom data. The correction relies on model fits to the energy spectra at each pixel in a given projection image. For analysis purposes, a numerical model of the Rollo phantom is employed to construct a projection image corresponding to that expected from an ideal imaging system. This is then convolved with the measured spatial point response (in air) of the actual imaging system to produce a realistic estimate of the scatter-free image. Summary statistics derived from comparisons between this image and the raw and scatter-corrected images are extremely reliable quantitative measures of the detrimental effects of scatter. Position and energy (xyE) list mode data were acquired for several isotopes, each with 0, 5, 10, and 15 cm of Lucite intervening between the phantom and the collimator face to induce varying amounts of scatter. Analysis of the /sup 99m/Tc and /sup 201/Tl data demonstrates a significant improvement in quantitation for high count images as well as for images of clinical count densities. >
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have