A measure of Anger-camera linearity: results with and without a corrector

A method for measuring the X and Y linearity of an Anger camera coupled to a computer is presented. It has similarities to, and differences from, the method recommended by the National Electrical Manufactures Association (NEMA). Test images are taken through a lead plate with parallel and equally spaced slots, and the locations of the lines in the images are fitted by least-squares with an equation that allows for slight misalignment. Discrepancies from the fit are calculated and displayed as a distribution over the camera field. The maximum and average discrepancies are tabulated. The field of view that is of interest is selectable within the analysis program. Among four large-field-of-view uncorrected cameras, three 37-tube types (two measured using Tc-99m and one with Au-195) show a similar degree of nonlinearity. However, the maximum and average discrepancies from linearity for a 61-tube prototype camera, measured using Tc-99m over a 15-in. field of view, are 40% of those for the other three. For the four cameras, an event-shifting on-line corrector with best-case sampling improves linearity by an average factor of 5.5, including both 15-and 11.25-in. fields of view.