Abstract
Abstract
PADME Run III searched for a resonant production of the X17 particle. The precise knowledge of the beam position is critical for the measurements of X17 and an array of 2 × 6 Timepix3 hybrid pixel detectors were used to measure the parameters of the incident positron beam. During the operation of the Timepix detector, some of the pixels reported missing data which introduced a challenge for the accurate estimation of the center position. A Neural Network Regression algorithm was developed and compared with the conventional Gaussian fitting technique over a simulated data with masked pixels.