A Post-Processing Algorithm to Correct Time Walk and Boost CTR to 100 ps Level

The coincident event analysis is of paramount importance in Positron Emission Tomography (PET). The result of this process, often termed Coincidence Time Resolution (CTR), is one of the most important quantitative factors that determines the performance of a PET system. Optimizing CTR, typically attempted by lowering threshold voltages in Leading Edge Discriminators (LEDs), presents a challenge due to prevalent pick-up noises. In light of this, in this paper, a post-processing algorithm is proposed. This algorithm is dedicated to a detector front-end with the addition of a Low Noise Amplifier (LNA). It can effectively identify outliers and tackle signal distortions so as to mitigate pick-up noises and finally improve CTR. The key contribution of this study is that it can notably improve CTR while still maintaining adequate detection efficiency. Extensive experiments are carried out to demonstrate that the proposed post-processing algorithm can effectively improve CTR, from about 240 ps down to around 100 ps, even with a crystal length of 20mm (the energy window is 450 keV to 600 keV). The power consumption of the single channel is only 0.12W.