Speaker
Description
We present a realistic 3D simulation workflow for pixel detectors, with emphasis on TCAD modeling of Timepix and CMS CROC pixel sensors and its connection to Monte Carlo-based synthetic data generation. Detailed electric-field and charge-transport simulations are used to model detector response, while Allpix Squared provides event-level Monte Carlo simulation and pixel-level observables such as charge, timing, position, and particle truth. These synthetic datasets can be used to study detector performance and to train spiking neural networks for tasks such as clustering, classification, and noise rejection. The work provides a pathway from sensor-level TCAD design studies to realistic training data for low-power edge processing in future pixel detector systems.