eduzhai > Applied Sciences > Engineering >

Fast and automated biomarker detection in breath samples with machine learning

  • Save

... pages left unread,continue reading

Document pages: 26 pages

Abstract: Volatile organic compounds (VOCs) in human breath can reveal a large spectrumof health conditions and can be used for fast, accurate and non-invasivediagnostics. Gas chromatography-mass spectrometry (GC-MS) is used to measureVOCs, but its application is limited by expert-driven data analysis that istime-consuming, subjective and may introduce errors. We propose a system toperform GC-MS data analysis that exploits deep learning pattern recognitionability to learn and automatically detect VOCs directly from raw data, thusbypassing expert-led processing. The new proposed approach showed to outperformthe expert-led analysis by detecting a significantly higher number of VOCs injust a fraction of time while maintaining high specificity. These resultssuggest that the proposed method can help the large-scale deployment ofbreath-based diagnosis by reducing time and cost, and increasing accuracy andconsistency.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×