We present the implementation of a practical iterative method for optical model estimation in Anger cameras for medical imaging. The optical model, also referred to as light response, represents a fundamental requirement for statistical event reconstruction in gamma detectors implementing monolithic scintillators. Despite we considered several approaches for the response calibration, based either on Monte Carlo optical modelling or experimental scanning of the detectors response, none of them was practical in producing a reliable optical model and, moreover, of practical use when the cameras are installed in the scanner. We therefore propose a more effective approach for optical model estimation. We designed and tested it on detectors based on Silicon PhotoMultipliers (SiPMs) and adopted in the INSERT project, aiming at the development of a non-rotational SPECT insert for MR scanners, composed by 10 detectors (preclinical version). The iterative method calibrates the light response starting from an experimental acquisition of a non-collimated source. Thus, the model is reliable, keeping the calibration procedure fast and simple. This permits to estimate, within the same session, the optical models for every gamma camera populating the SPECT in few tens of minutes. The estimated optical model feeds a Maximum Likelihood algorithm for optimal event reconstruction, providing better uniformity and linearity together with recovery of the field of view. We tested the reconstructed images quality by acquiring irradiation profiles with a gamma-camera designed for the preclinical version of the INSERT SPECT. The maximum deviation from linearity results inferior to 1 mm. Deviation from uniformity is less than the 15 % of the average number of counts. The optical model is also well adaptable to the variability of detection response thanks to the well uniform behavior of the SiPMs adopted.
Light response estimation and gamma events reconstruction in gamma-detectors based on continuous scintillators and SiPMs
Occhipinti, M.;Busca, P.;Fiorini, C.
2017-01-01
Abstract
We present the implementation of a practical iterative method for optical model estimation in Anger cameras for medical imaging. The optical model, also referred to as light response, represents a fundamental requirement for statistical event reconstruction in gamma detectors implementing monolithic scintillators. Despite we considered several approaches for the response calibration, based either on Monte Carlo optical modelling or experimental scanning of the detectors response, none of them was practical in producing a reliable optical model and, moreover, of practical use when the cameras are installed in the scanner. We therefore propose a more effective approach for optical model estimation. We designed and tested it on detectors based on Silicon PhotoMultipliers (SiPMs) and adopted in the INSERT project, aiming at the development of a non-rotational SPECT insert for MR scanners, composed by 10 detectors (preclinical version). The iterative method calibrates the light response starting from an experimental acquisition of a non-collimated source. Thus, the model is reliable, keeping the calibration procedure fast and simple. This permits to estimate, within the same session, the optical models for every gamma camera populating the SPECT in few tens of minutes. The estimated optical model feeds a Maximum Likelihood algorithm for optimal event reconstruction, providing better uniformity and linearity together with recovery of the field of view. We tested the reconstructed images quality by acquiring irradiation profiles with a gamma-camera designed for the preclinical version of the INSERT SPECT. The maximum deviation from linearity results inferior to 1 mm. Deviation from uniformity is less than the 15 % of the average number of counts. The optical model is also well adaptable to the variability of detection response thanks to the well uniform behavior of the SiPMs adopted.File | Dimensione | Formato | |
---|---|---|---|
08069405.pdf
Accesso riservato
:
Publisher’s version
Dimensione
1.02 MB
Formato
Adobe PDF
|
1.02 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.