Precise real-time correction of Anger camera deadtime losses

An earlier paper dealt with modeling of the camera in terms of the resolving times, τ0 and T, of the paralyzable detector and nonparalyzable computer system, respectively, for the case of a full energy window. A second paper presented a decaying source method for the accurate real-time measurement of these resolving times. The present paper first shows that the detector system can be treated as a single device with a resolving time τ0 dependent on source distribution. It then discusses camera operation with an energy window, window fraction being fw=Rp/Rd≤1, where Rd and Rp are the detector and pulse-height-analyzer (PHA) outputs, respectively. The detector resolving time is shown to vary with window fraction according to τ0p=τ0p/fw, while T is unaffected, so that operation may be paralyzable or nonparalyzable depending on window setting and the ratio kT=T/τ0. Regions of interest are described in terms of the ROI fraction, fr=Rr/R≤1, and resolving time, τ0r=τ0p/fr, where R and Rr are the recorded count rates for the field-of-view and the region-of-interest, respectively. As τ0p and τ0r are expected to vary with input rate, it is shown that these can be measured in real-time using the decaying source method. It is then shown that camera operation both with fw≤1 and fr≤1 can be described by the simple paralyzable equation r=ne-n, where n=Nwτ0p=Nrτ0r and r=Rpτ0p=Rrτ0r, Nw, and Nr being the input rates within the energy window and the region of interest, respectively. An analytical solution to the paralyzable equation is then presented, which enables the input rates Nw=n/τ0p and Nr=n/τ0r to be obtained correct to better than 0.52% all the way up to the peak response point of the camera