Refinement of spectra using a deep neural network: Fully automated removal of noise and background
Standard
Refinement of spectra using a deep neural network: Fully automated removal of noise and background. / Gebrekidan, Medhanie Tesfay; Knipfer, Christian; Braeuer, Andreas Siegfried.
in: J RAMAN SPECTROSC, Jahrgang 52, Nr. 3, 09.03.2021, S. 723-736.Publikationen: SCORING: Beitrag in Fachzeitschrift/Zeitung › SCORING: Zeitschriftenaufsatz › Forschung › Begutachtung
Harvard
APA
Vancouver
Bibtex
}
RIS
TY - JOUR
T1 - Refinement of spectra using a deep neural network: Fully automated removal of noise and background
AU - Gebrekidan, Medhanie Tesfay
AU - Knipfer, Christian
AU - Braeuer, Andreas Siegfried
N1 - Funding Information: The project leading to this result has received funding from the European Union's Horizon 2020 research and innovation program under ERC Starting Grant agreement 637654 (Inhomogeneities). It has also received funding from the Wilhelm Sander‐Stiftung, Munich, Germany (Grants 2017.111.1 and 2017.111.2). Publisher Copyright: © 2021 The Authors. Journal of Raman Spectroscopy published by John Wiley & Sons Ltd. Copyright: Copyright 2021 Elsevier B.V., All rights reserved.
PY - 2021/3/9
Y1 - 2021/3/9
N2 - We report the potential of U-Net deep neural network for the efficient removal of noise and background from raw Raman spectra. The U-Net method was first trained on simulated spectra and then tested with experimental spectra. The quality of the test results was quantified via different signal-to-noise ratios and the structural similarity index metric. The U-Net recovered Raman spectra feature a high structural similarity index, even for raw spectra that were dominated by background. The U-Net model does not rely on any human intervention.
AB - We report the potential of U-Net deep neural network for the efficient removal of noise and background from raw Raman spectra. The U-Net method was first trained on simulated spectra and then tested with experimental spectra. The quality of the test results was quantified via different signal-to-noise ratios and the structural similarity index metric. The U-Net recovered Raman spectra feature a high structural similarity index, even for raw spectra that were dominated by background. The U-Net model does not rely on any human intervention.
KW - deep learning
KW - fluorescence rejection
KW - noise reduction
KW - Raman spectra
KW - U-Net
U2 - 10.1002/jrs.6053
DO - 10.1002/jrs.6053
M3 - SCORING: Journal article
AN - SCOPUS:85099266182
VL - 52
SP - 723
EP - 736
JO - J RAMAN SPECTROSC
JF - J RAMAN SPECTROSC
SN - 0377-0486
IS - 3
ER -