Speaker
Description
Reliable particle identification (PID) is essential for any high energy physics experiment. The Time-Of-Propagation (TOP) counter is a ring-imaging Cherenkov detector designed to identify charged particles in the barrel region of the Belle II detector. When charged particles cross a TOP quartz bar, they emit Cherenkov photons, which are detected by a pixelated photodetector at one end of the bar. Different charged particles with the same momentum and angle of impact on the bar will have different patterns of time of photon arrival versus position of photon impact on the detector plane. Currently, PID information is obtained by comparing the measured pattern with analytically calculated probability density functions for different charged particle hypotheses. We aim to improve the PID performance of the TOP by using machine learning to recognize and distinguish between the different patterns.
