The Estimation and Removal of Unknown Background Signatures in Raman-Shift Spectra via SWTPE

摘要

We present a technique, called the sliding-window technique for pedestal estimation (SWTPE), for estimation and removal of unknown arbitrarily-varying noisy background (pedestal) from Raman-shift spectra that may contain Raman-signatures of chemical agents. ROC plots depicting Pd驴0.8 vs. Pfa 驴0.1 are presented for two example backgrounds to demonstrate the performance of the algorithm.