Premium
Efficiency of visual peak detection in X‐ray photoelectron spectra
Author(s) -
Suzuki Mineharu,
Fukushima Sei,
Tanuma Shigeo
Publication year - 2008
Publication title -
surface and interface analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.52
H-Index - 90
eISSN - 1096-9918
pISSN - 0142-2421
DOI - 10.1002/sia.2894
Subject(s) - spectral line , principal component analysis , logarithm , analytical chemistry (journal) , chemistry , pattern recognition (psychology) , artificial intelligence , mathematics , computer science , physics , chromatography , mathematical analysis , astronomy
We report initial results of a VAMAS/TWA2 project to evaluate procedures for automated peak detection in X‐ray photoelectron spectra. As a reference for investigations of the efficiency of automated peak‐detection software, we report the efficiency of visual peak detection in three test spectra. It was found that (i) characteristics of analysts are grouped into four categories using principal component analysis (PCA); the first participant group to detect large numbers of peaks for the three test spectra, the second one to detect small numbers of peaks for them, the third one to detect similar numbers of peaks, and the fourth one to detect a relatively large number of peaks for one of them and small numbers for two of them, (ii) scattering of detected peak numbers seems to depend on detection of medium‐intensity peaks because of participants' subjectivity or ambivalence for judgment of intensity, and (iii) the peaks that are detected by the analysts with a detectability more than 75% almost correspond to the peak signal‐to‐noise(S/N) ratio of more than 10 in logarithmic expression. Copyright © 2008 John Wiley & Sons, Ltd.