Speaker
Dr
Eiji Yoshida
(National Institute of Radiological Sciences)
Description
In a conventional PET detector, detected events are projected to 2D position
histogram with Anger calculation for a crystal identification. However, the measured
histogram is affected by inter-crystal scatterings (ICS) which occur in a whole
detector. Peaks with projected each crystal in the histogram are blurred, and it
causes ICS’s mispositioning. A depth of interaction (DOI) detector has been developed
for the small animal PET scanner: jPET-RD. This DOI detector uses 32x32 crystals with
four layers and a 256-channel multi-anode flat panel photomultiplier tube (FP-PMT)
was developed by Hamamatsu Photonics K.K. Each crystal element is 1.45x1.45x4.5 mm.
The FP-PMT has a large useful area (49x49 mm) and small anode pitch (3.04 mm).
Therefore, the FP-PMT can extensively trace the behavior of incident gamma rays in
the crystals including ICS. We therefore propose a novel method for ICS estimation
using a statistical pattern recognition based on the support vector machine (SVM). In
this study, we applied the SVM for discriminating photoelectric events from ICS
events generated from multiple-anode outputs. The SVM was trained by uniform
irradiation events generated from a detector simulator using Monte Carlo calculation.
Initial result of ICS identification is about 80 % for non-training data. This method
can archive a true subtraction of ICS from measured events, and it is also useful for
random correction in PET.
Author
Dr
Eiji Yoshida
(National Institute of Radiological Sciences)
Co-authors
Dr
Fumihiko Nishikido
(Waseda University)
Dr
Hideo Murayama
(National Institute of Radiological Sciences)
Dr
Keishi Kitamura
(Shimadzu Co.)
Dr
Kengo Shibuya
(National Institute of Radiological Sciences)
Dr
Taiga Yamaya
(National Institute of Radiological Sciences)
Dr
Yuichi Kimura
(Tokyo Metropolitan Institute of Gerontology)