Speaker
Dr
Ioannis Kaissas
(Greek Atomic Energy Commission)
Description
In this work we compare the traditional correlation process of a Coded Aperture device to estimate the spatial coordinates of γ-emitters with a different approach: We have developed machine learning algorithms based on Gradient Boosted Decision Trees (BDTG) and Deep Neural Networks (DNN). The algorithms have been trained using 18000 shadowgrams created with simulation. A custom fast simulation tool was used to produce shadowgrams due to sources placed randomly at 18000 different positions within the FOV and up to a distance of 4.5m from the detector plane. The performance of the algorithms has been evaluated with the aid of a different independent sample of shadowgrams.
Primary authors
Dr
Konstantinos Karafasoulis
(Hellenic Army Academy)
Dr
Ioannis Kaissas
(Greek Atomic Energy Commission)
Dr
Christos Papadimitropoulos
(Department of Aerospace Science and Technology, National and Kapodistrian University of Athens)
Dr
Constantinos Potiriadis
(Greek Atomic Energy Commission)
Prof.
Charalambos Pan Lambropoulos
(Department of Aerospace Science and Technology, National and Kapodistrian University of Athens)