Waveform discrimination by fitting derivative of synchronized ideal normalized curves, dSINC fit

TL;DR

dSINC introduces a novel waveform discrimination algorithm that fits the derivative of waveforms to ideal models, effectively reducing noise sensitivity and peak piling issues in multi-layer scintillator measurements.

Contribution

The paper presents a new fitting-based approach for waveform discrimination that eliminates the need for traditional feature extraction, improving robustness against noise.

Findings

Effective discrimination in noisy conditions

Reduced peak piling sensitivity

Improved accuracy over traditional methods

Abstract

dSINC proposes an alternative algorithm for waveform discrimination of measurement data from multi-layer scintillator sandwich designs. dSINC attempts to solve problems related to noise and peaks-piling sensitivity in the feature extraction step of traditional KNN waveform discrimination, by fitting the derivative of the entire gain section of the waveform against ideal waveforms learned from training data and thereby completely sidestepping the problems of feature extraction.