Simple performance measures for imaging system optimization [SPECT application]

The goal of this work was to develop a scalar performance index for obtaining a quick look at the performance potential of different imaging system geometries-especially those where the system model is incompletely known. The Fisher information for a single “signal” parameter was investigated for this purpose, but was found to be insufficient when the task was optimizing the resolution-sensitivity tradeoff of a parallel hole collimator. On the other hand the 2×2 Cramer-Rao bound corresponding to the estimation of the intensities of appropriately chosen signal and background distributions may be useful for initial imaging system evaluations. However, caution must be used; results are quite sensitive to the signal and background distributions