Abstract

Analyses of the composition of exhaled gases are among important areas of non-invasive medicine. Diagnostics based on the analysis of exhaled gases offers several advantages over traditional laboratory methods. An analysis of a gas mixture is safe for the staff, because it does not involve chemical and biological fluids. The analysis is relatively inexpensive, takes little time, and ensures the detection of volatile components in the exhaled gas at the trace level of substances in real time. This paper discusses algorithms for the processing of mass spectra of exhaled gases obtained on an MC7-200 quadrupole mass spectrometer with electron ionization and the direct capillary injection of a sample. A comparative evaluation of algorithms for the detection of spectral lines in mass spectra based on data convolution with a model of response function or its derivatives is given. The Hermite function, the Gaussian function, and its even-order derivatives are considered as convolution kernels. The results of evaluation of the potential for increasing the signal-to-noise ratio by algorithms of convolution with these functions are presented.

中文翻译：

---

基于相关算法的呼出气体质谱处理

摘要

呼出气体的成分分析是非侵入性医学的重要领域。与传统的实验室方法相比，基于呼出气体分析的诊断具有许多优势。对气体混合物进行分析对于工作人员而言是安全的，因为它不涉及化学和生物流体。该分析相对便宜，花费的时间很少，并且可以确保实时检测痕量物质中呼出气体中的挥发性成分。本文讨论了用电子电离和样品的直接毛细管注入在MC7-200四极质谱仪上获得的呼出气体质谱图的处理算法。给出了基于响应函数或其导数模型的数据卷积的质谱图中谱线检测算法的比较评估。Hermite函数，Gaussian函数及其偶数阶导数被视为卷积核。提出了通过与这些函数卷积的算法对提高信噪比潜力的评估结果。