DocumentCode
2624618
Title
Signal deconvolution concept combined with Cubic Spline interpolation to improve timing with phoswich pet detectors
Author
Semmaoui, Hicham ; Tétrault, Marc-André ; Lecomte, Roger ; Fontaine, Réjean
Author_Institution
Department of Electrical and Computer Engineering, Université de Sherbrooke, QC J1K 2R1, Canada
fYear
2008
fDate
19-25 Oct. 2008
Firstpage
2061
Lastpage
2067
Abstract
PET imaging scanners based on all-digital architecture offer greater data processing flexibility and the possibility to recalibrate the system with simple software procedures. The LabPET™ scanner, designed at the Université de Sherbrooke, is one such device. It is built around dual LYSO/LGSO scintillators in a phoswich arrangement coupled to Avalanche Photodiodes (APD)and combined with highly parallel readout and processing electronics. This approach enables the implementation of advanced real time digital signal processing methods to compute energy resolution, crystal identification and the arrival time of events. Timing extraction represents the highest challenge because of the low sampling frequency (45 MHz), the quantization error and the presence of a Zero Order Hold (ZOH) in the system. The aim of this paper is to present a method to increase coincidence time accuracy by adequately addressing each of these limitations. The proposed method uses a Deconvolution concept based on adaptive filter theory preceded by Cubic Spline interpolation to improve digital timing performance. The method achieves 4.5 ns, 8.2 ns and 6.5 ns timing resolution with, LYSO-LYSO, LGSO-LGSO and mixed crystals phoswich coincidences, respectively.
Keywords
Avalanche photodiodes; Computer architecture; Data processing; Deconvolution; Detectors; Interpolation; Positron emission tomography; Software systems; Spline; Timing; Adaptive Filter; Arrival time; Crystal Identification; Cubic Spline; PET scanner; Signal deconvolution;
fLanguage
English
Publisher
ieee
Conference_Titel
Nuclear Science Symposium Conference Record, 2008. NSS '08. IEEE
Conference_Location
Dresden, Germany
ISSN
1095-7863
Print_ISBN
978-1-4244-2714-7
Electronic_ISBN
1095-7863
Type
conf
DOI
10.1109/NSSMIC.2008.4774887
Filename
4774887
Link To Document