Observer performance as a measure of goodness in the evaluation of scintigraphic image processing

Evaluation by human observers seems to be the only really viable method of establishing the worth of the many types of radionuclide image processing. Someday links between mathematical formulations and human performance will become established. When those links are available, the effect of parametric variation in system processing details may be more readily evaluated in a meaningful way, but until these are available, we are constrained and in fact, obligated to using real observers evaluating numerous examples of representative scans