A Singles-Rate Based Dual Dead-Time Constant Model for Dead-Time Correction

Dead-time effects significantly degrade the quantitative accuracy of positron emission tomography (PET) imaging under high-activity conditions. In PET systems, dead-time refers to the interval after a gamma event is detected during which the system is unable to register subsequent events. Although these effects fundamentally occur at the singles detection level, most of current correction methods are developed at the coincidences level. With the advent of all-digital PET systems, access to complete singles data now allows for correction strategies to be developed directly at the singles level. In this study, we propose a singles-based dual dead-time constant model, separately accounting for detector-level (non-paralyzable) and electronics-level (paralyzable) dead-time. The detector component is estimated from the time intervals between consecutive singles, while the electronics component is fitted by comparing the observed singles rate with the ideal rate extrapolated from low-activity measurements. Validation experiments using an image quality (IQ) phantom on a pre-clinical PET/CT scanner demonstrate that our approach achieves relative count rate errors within ±2.5% at activity levels up to 60 MBq. These results demonstrate the feasibility of the proposed singles-based approach under high count-rate conditions and highlight its potential as a fundamentally different strategy for dead-time correction in PET imaging.