New preprint: omnidirectional gamma-ray detection with machine learning
A compact GAGG detector with perpendicular SiPM readout and machine learning to estimate source distance
Our latest preprint, Omnidirectional Radiation Detector with Perpendicular Dual Silicon Photomultiplier Readout – Directional Sensitivity and Machine Learning Source Positioning, is now available on arXiv.
The study presents a compact detector design for gamma-ray detection over the full solid angle (4π). It combines 64 GAGG scintillation crystals in a 4 × 4 × 4 matrix with silicon photomultiplier readout on two perpendicular faces.
Using Geant4 simulations, we investigated detection efficiency and directional sensitivity. We also trained and validated XGBoost machine learning models on simulated data, successfully estimating source-to-detector distances at gamma-ray energies of 511, 662 and 1275 keV.
These estimates could provide initial spatial information for Compton image reconstruction, helping reduce reconstruction time and computational requirements.
The work was carried out within the VMDScan project, funded by the European Union – NextGenerationEU.
Authors: Ana Marija Kožuljević, Gabriela Jazvac, Luka Lotina and Luka Pavelić.