Correcting Nonuniform Light Distribution: Implementation of ANFIS on a Gamma Position-Sensitive Module

Nuclear imaging techniques rely on the accurate positioning of events to produce the final images. Among the factors affecting the event positioning accuracy is the nonuniform light distribution between different regions of gamma detectors. In this article, we attempt to implement the adaptive neuro-fuzzy inference system (ANFIS) as a scintillation event positioning algorithm to improve the linearity and uniformity of the response. Therefore, we trained the network using signal data collected from a monolithic scintillator coupled to a position-sensitive photomultiplier tube (PS-PMT). We evaluated the position detection performance of the system after convergence with a test dataset. The results showed a 1.149-mm RMSE and over 98%
$R^{2}$
value for position estimation. Moreover, the average full-width at half-maximum (FWHM) values obtained by ANFIS were 1.18 and 1.57 mm for the
$x$
- and
$y$
-axes, respectively. Therefore, the implementation of ANFIS as a position estimation algorithm can minimize the limitations of the edge distortion effect in the final results.