Learning SPECT detector response for Monte-Carlo simulations

D. Sarrut (1) , Nils Krah (2) , J N Badel (1) , Jean Michel Létang (1)

1 Imagerie Tomographique et Radiothérapie
2 IPNL - Institut de Physique Nucléaire de Lyon

The use of Artificial Neural Networks (ANN) was investigated to learn detector response in Monte-Carlo (MC) simulation of imaging systems in order to replace computationally costly particle tracking in the detector. The proposed method is applied to speed-up simulation of collimator-detector systems with Angular Response Function (ARF) for SPECT imaging.

Domaines: Modélisation et simulation; Imagerie médicale

D. Sarrut, Nils Krah, J N Badel, Jean Michel Létang. Learning SPECT detector response for Monte-Carlo simulations. Third Geant4 International User Conference, Oct 2018, Bordeaux, France. ⟨hal-01926625⟩