RECURSIVE CALCULATION OF THE 2-DIMENSIONAL MAXIMUM-LIKELIHOOD POSITION ESTIMATE FOR A SCINTILLATION CAMERA

Maximum likelihood position estimation of a scintillation event in a gamma-ray imaging detector has been described using two independent one-dimensional estimates in x and y. This approximation is only valid when the sampled light-spread function is separable, and spatial distortion is observed when this constraint is violated. We have investigated recursive calculation of the 2-D position estimate using a sequence of one-dimensional estimates. Mean phototube response functions were computed from Monte Carlo simulations of the light-spread function, and position was estimated for scintillation events using recursive estimation. After 3 steps the maximum errors were ∣δx∣ = 0.04mm and ∣δy∣ = 0.06mm for the case of noise-free data and when mean response functions were known at each estimated location. Using response functions interpolated over 6 mm intervals gives increased distortion with average x and y errors of 0.097 and 0.082 mm respectively. The maximum errors are ∣δx∣= 0.367mm and ∣δy∣ = 0.308mm. Recursive one-dimensional estimates of the most likely 2-D location of a scintillation event converge rapidly to within a fraction of a millimeter of the true location. Copyright © 1990 by The Institute of Electrical and Electronics Engineers, Inc.