Clustering of Alpha Curves in Differential Mobility Spectrometry Data

Abstract

Volatile organic compounds (VOC) can be analyzed<br/>and classified based on dispersion plots yielded by Differential<br/>Mobility Spectrometry (DMS). These dispersion plots contain<br/>traces, also known as alpha curves that indicate the presence of<br/>VOCs with specific charge and cross-sectional area. However,<br/>often the dispersion plots are analyzed without utilizing this<br/>information but rather using the pixel values independently or<br/>analyzing values at a specific high electric field level. This paper<br/>proposes a technique for extracting and clustering alpha curves<br/>from dispersion plots. The cluster information could then be<br/>used inside existing detection and classification algorithms to<br/>potentially improve their accuracy