Improved detection of chemical substances from colorimetric sensor data using probabilistic machine learning

Lasse Lohilahti Mølgaard, Ole Thomsen Buus, Jan Larsen, Hamid Babamoradi, Ida Lysgaard Thygesen, Milan Laustsen, Jens Kristian Munk, Eleftheria Dossi, Caroline O'Keeffe, Lina Lässig, Sol Tatlow, Lars Sandström, Mogens Havsteen Jakobsen

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

422 Downloads (Pure)

Abstract

We present a data-driven machine learning approach to detect drug- and explosives-precursors using colorimetric sensor technology for air-sampling. The sensing technology has been developed in the context of the CRIM-TRACK project. At present a fully- integrated portable prototype for air sampling with disposable sensing chips and automated data acquisition has been developed. The prototype allows for fast, user-friendly sampling, which has made it possible to produce large datasets of colorimetric data for different target analytes in laboratory and simulated real-world application scenarios. To make use of the highly multi-variate data produced from the colorimetric chip a number of machine learning techniques are employed to provide reliable classification of target analytes from confounders found in the air streams. We demonstrate that a data-driven machine learning method using dimensionality reduction in combination with a probabilistic classifier makes it possible to produce informative features and a high detection rate of analytes. Furthermore, the probabilistic machine learning approach provides a means of automatically identifying unreliable measurements that could produce false predictions. The robustness of the colorimetric sensor has been evaluated in a series of experiments focusing on the amphetamine pre-cursor phenylacetone as well as the improvised explosives pre-cursor hydrogen peroxide. The analysis demonstrates that the system is able to detect analytes in clean air and mixed with substances that occur naturally in real-world sampling scenarios. The technology under development in CRIM-TRACK has the potential as an effective tool to control trafficking of illegal drugs, explosive detection, or in other law enforcement applications.
Original languageEnglish
Title of host publicationProceedings of SPIE
Number of pages8
Volume10183
PublisherSPIE - International Society for Optical Engineering
Publication date2017
Article number1018307
ISBN (Print)9781510608672
DOIs
Publication statusPublished - 2017
EventChemical, Biological, Radiological, Nuclear, and Explosives (CBRNE) Sensing XVIII - Anaheim Convention Center, Anaheim, United States
Duration: 9 Apr 201713 Apr 2017

Conference

ConferenceChemical, Biological, Radiological, Nuclear, and Explosives (CBRNE) Sensing XVIII
LocationAnaheim Convention Center
CountryUnited States
CityAnaheim
Period09/04/201713/04/2017
SeriesProceedings of S P I E - International Society for Optical Engineering
Volume10183
ISSN0277-786X

Bibliographical note

Copyright 2017 Society of Photo Optical Instrumentation Engineers. One print or electronic copy may be made for personal use only. Systematic reproduction and distribution, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper are prohibited.

Keywords

  • Artificial nose
  • Colorimetric sensor array
  • Machine learning

Cite this

Mølgaard, L. L., Buus, O. T., Larsen, J., Babamoradi, H., Thygesen, I. L., Laustsen, M., ... Jakobsen, M. H. (2017). Improved detection of chemical substances from colorimetric sensor data using probabilistic machine learning. In Proceedings of SPIE (Vol. 10183). [1018307] SPIE - International Society for Optical Engineering. Proceedings of S P I E - International Society for Optical Engineering, Vol.. 10183 https://doi.org/10.1117/12.2262468
Mølgaard, Lasse Lohilahti ; Buus, Ole Thomsen ; Larsen, Jan ; Babamoradi, Hamid ; Thygesen, Ida Lysgaard ; Laustsen, Milan ; Munk, Jens Kristian ; Dossi, Eleftheria ; O'Keeffe, Caroline ; Lässig, Lina ; Tatlow, Sol ; Sandström, Lars ; Jakobsen, Mogens Havsteen. / Improved detection of chemical substances from colorimetric sensor data using probabilistic machine learning. Proceedings of SPIE. Vol. 10183 SPIE - International Society for Optical Engineering, 2017. (Proceedings of S P I E - International Society for Optical Engineering, Vol. 10183).
@inproceedings{6733ddb0d0364c4cbfd3d1fa32b26b0c,
title = "Improved detection of chemical substances from colorimetric sensor data using probabilistic machine learning",
abstract = "We present a data-driven machine learning approach to detect drug- and explosives-precursors using colorimetric sensor technology for air-sampling. The sensing technology has been developed in the context of the CRIM-TRACK project. At present a fully- integrated portable prototype for air sampling with disposable sensing chips and automated data acquisition has been developed. The prototype allows for fast, user-friendly sampling, which has made it possible to produce large datasets of colorimetric data for different target analytes in laboratory and simulated real-world application scenarios. To make use of the highly multi-variate data produced from the colorimetric chip a number of machine learning techniques are employed to provide reliable classification of target analytes from confounders found in the air streams. We demonstrate that a data-driven machine learning method using dimensionality reduction in combination with a probabilistic classifier makes it possible to produce informative features and a high detection rate of analytes. Furthermore, the probabilistic machine learning approach provides a means of automatically identifying unreliable measurements that could produce false predictions. The robustness of the colorimetric sensor has been evaluated in a series of experiments focusing on the amphetamine pre-cursor phenylacetone as well as the improvised explosives pre-cursor hydrogen peroxide. The analysis demonstrates that the system is able to detect analytes in clean air and mixed with substances that occur naturally in real-world sampling scenarios. The technology under development in CRIM-TRACK has the potential as an effective tool to control trafficking of illegal drugs, explosive detection, or in other law enforcement applications.",
keywords = "Artificial nose, Colorimetric sensor array, Machine learning",
author = "M{\o}lgaard, {Lasse Lohilahti} and Buus, {Ole Thomsen} and Jan Larsen and Hamid Babamoradi and Thygesen, {Ida Lysgaard} and Milan Laustsen and Munk, {Jens Kristian} and Eleftheria Dossi and Caroline O'Keeffe and Lina Lässig and Sol Tatlow and Lars Sandström and Jakobsen, {Mogens Havsteen}",
note = "Copyright 2017 Society of Photo Optical Instrumentation Engineers. One print or electronic copy may be made for personal use only. Systematic reproduction and distribution, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper are prohibited.",
year = "2017",
doi = "10.1117/12.2262468",
language = "English",
isbn = "9781510608672",
volume = "10183",
series = "Proceedings of S P I E - International Society for Optical Engineering",
publisher = "SPIE - International Society for Optical Engineering",
booktitle = "Proceedings of SPIE",

}

Mølgaard, LL, Buus, OT, Larsen, J, Babamoradi, H, Thygesen, IL, Laustsen, M, Munk, JK, Dossi, E, O'Keeffe, C, Lässig, L, Tatlow, S, Sandström, L & Jakobsen, MH 2017, Improved detection of chemical substances from colorimetric sensor data using probabilistic machine learning. in Proceedings of SPIE. vol. 10183, 1018307, SPIE - International Society for Optical Engineering, Proceedings of S P I E - International Society for Optical Engineering, vol. 10183, Chemical, Biological, Radiological, Nuclear, and Explosives (CBRNE) Sensing XVIII, Anaheim, United States, 09/04/2017. https://doi.org/10.1117/12.2262468

Improved detection of chemical substances from colorimetric sensor data using probabilistic machine learning. / Mølgaard, Lasse Lohilahti; Buus, Ole Thomsen; Larsen, Jan; Babamoradi, Hamid; Thygesen, Ida Lysgaard; Laustsen, Milan; Munk, Jens Kristian; Dossi, Eleftheria ; O'Keeffe, Caroline; Lässig, Lina; Tatlow, Sol; Sandström, Lars; Jakobsen, Mogens Havsteen.

Proceedings of SPIE. Vol. 10183 SPIE - International Society for Optical Engineering, 2017. 1018307 (Proceedings of S P I E - International Society for Optical Engineering, Vol. 10183).

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

TY - GEN

T1 - Improved detection of chemical substances from colorimetric sensor data using probabilistic machine learning

AU - Mølgaard, Lasse Lohilahti

AU - Buus, Ole Thomsen

AU - Larsen, Jan

AU - Babamoradi, Hamid

AU - Thygesen, Ida Lysgaard

AU - Laustsen, Milan

AU - Munk, Jens Kristian

AU - Dossi, Eleftheria

AU - O'Keeffe, Caroline

AU - Lässig, Lina

AU - Tatlow, Sol

AU - Sandström, Lars

AU - Jakobsen, Mogens Havsteen

N1 - Copyright 2017 Society of Photo Optical Instrumentation Engineers. One print or electronic copy may be made for personal use only. Systematic reproduction and distribution, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper are prohibited.

PY - 2017

Y1 - 2017

N2 - We present a data-driven machine learning approach to detect drug- and explosives-precursors using colorimetric sensor technology for air-sampling. The sensing technology has been developed in the context of the CRIM-TRACK project. At present a fully- integrated portable prototype for air sampling with disposable sensing chips and automated data acquisition has been developed. The prototype allows for fast, user-friendly sampling, which has made it possible to produce large datasets of colorimetric data for different target analytes in laboratory and simulated real-world application scenarios. To make use of the highly multi-variate data produced from the colorimetric chip a number of machine learning techniques are employed to provide reliable classification of target analytes from confounders found in the air streams. We demonstrate that a data-driven machine learning method using dimensionality reduction in combination with a probabilistic classifier makes it possible to produce informative features and a high detection rate of analytes. Furthermore, the probabilistic machine learning approach provides a means of automatically identifying unreliable measurements that could produce false predictions. The robustness of the colorimetric sensor has been evaluated in a series of experiments focusing on the amphetamine pre-cursor phenylacetone as well as the improvised explosives pre-cursor hydrogen peroxide. The analysis demonstrates that the system is able to detect analytes in clean air and mixed with substances that occur naturally in real-world sampling scenarios. The technology under development in CRIM-TRACK has the potential as an effective tool to control trafficking of illegal drugs, explosive detection, or in other law enforcement applications.

AB - We present a data-driven machine learning approach to detect drug- and explosives-precursors using colorimetric sensor technology for air-sampling. The sensing technology has been developed in the context of the CRIM-TRACK project. At present a fully- integrated portable prototype for air sampling with disposable sensing chips and automated data acquisition has been developed. The prototype allows for fast, user-friendly sampling, which has made it possible to produce large datasets of colorimetric data for different target analytes in laboratory and simulated real-world application scenarios. To make use of the highly multi-variate data produced from the colorimetric chip a number of machine learning techniques are employed to provide reliable classification of target analytes from confounders found in the air streams. We demonstrate that a data-driven machine learning method using dimensionality reduction in combination with a probabilistic classifier makes it possible to produce informative features and a high detection rate of analytes. Furthermore, the probabilistic machine learning approach provides a means of automatically identifying unreliable measurements that could produce false predictions. The robustness of the colorimetric sensor has been evaluated in a series of experiments focusing on the amphetamine pre-cursor phenylacetone as well as the improvised explosives pre-cursor hydrogen peroxide. The analysis demonstrates that the system is able to detect analytes in clean air and mixed with substances that occur naturally in real-world sampling scenarios. The technology under development in CRIM-TRACK has the potential as an effective tool to control trafficking of illegal drugs, explosive detection, or in other law enforcement applications.

KW - Artificial nose

KW - Colorimetric sensor array

KW - Machine learning

U2 - 10.1117/12.2262468

DO - 10.1117/12.2262468

M3 - Article in proceedings

SN - 9781510608672

VL - 10183

T3 - Proceedings of S P I E - International Society for Optical Engineering

BT - Proceedings of SPIE

PB - SPIE - International Society for Optical Engineering

ER -

Mølgaard LL, Buus OT, Larsen J, Babamoradi H, Thygesen IL, Laustsen M et al. Improved detection of chemical substances from colorimetric sensor data using probabilistic machine learning. In Proceedings of SPIE. Vol. 10183. SPIE - International Society for Optical Engineering. 2017. 1018307. (Proceedings of S P I E - International Society for Optical Engineering, Vol. 10183). https://doi.org/10.1117/12.2262468