We use the power density spectra obtained by fluctuation-enhanced sensing of bacterial odors (Escherichia coli and Anthrax-surrogate Bacillus subtilis) to generate new, highly distinguishable, types of patterns based on the average slope of the spectra in different frequency ranges. Such plots can be considered as "fingerprints" of bacterial odors. Three different ways of pattern generation are tested, including a simple binary version. The obtained patterns are simple enough to identify the situation by the naked eye without a pattern recognizer.
Cite
@article{arxiv.0902.1368,
title = {Identification of Bacteria by Patterns Generated from Odor Spectra},
author = {Hung-Chih Chang and Laszlo B. Kish and Maria D. King and Chiman Kwan},
journal= {arXiv preprint arXiv:0902.1368},
year = {2012}
}