Classification of colorimetric sensor data using time series

Abstract

Colorimetric sensors are widely used as pH indicators, medical diagnostic devices and detection devices. The colorimetric sensor captures the color changes of a chromic chemical (dye) or array of chromic chemicals when exposed to a target substance (analyte). Sensing is typically carried out using the difference in dye color before and after exposure. This approach neglects the kinetic response, that is, the temporal evolution of the dye, which potentially contains additional information. We investigate the importance of the kinetic response by collecting a sequence of images over time. We applied end-to-end learning using three different convolution neural networks (CNN) and a recurrent network. We compared the performance to logistic regression, k-nearest-neighbor and random forest, where these methods only use the difference color from start to end as feature vector. We found that the CNNs were able to extract features from the kinetic response profiles that significantly improves the accuracy of the sensor. Thus, we conclude that the kinetic responses indeed improves the accuracy, which paves the way for new and better chemical sensors based on colorimetric responses.

Machine learning methods for the detection of explosives, drugs and precursor chemicals gathered using a colorimetric sniffer sensor

“…Recently, time series data has shown helpful in improving the detection capability of several explosives and their precursors. 15 This kind of data is helpful in understanding the temporal evolution of the dye color changes as the reaction progresses. Table S2 † shows the number of measurements of each analyte present in each dataset.…”

“…The results obtained indicate that the kinetic nature of the data is useful in improving the detection capability of explosives and precursors as shown by previous work. 15 However, the use of deep learning models requires access to large amounts of data measurements. Thus, the use of such models is limited to the case where we have a sufficient number of data instances.…”

“…The implementation of a colorimetric sensor system in the portable CRIM-TRACK device has ensured the detection of trace amounts of substances (i.e., HMEs), at parts-per-trillion (ppt) level in the air, with near real-time detection capability and no contact operator-threat. 11,15 It incorporates similar sensitivity and portability of sniffer dogs with the discriminatory ability of instrumentation without the use of sample preparation or a highly trained operator.…”

“…In the context of machine learning methods for the detection of explosives and precursors, previous works have used colorimetric sensor data with machine learning classiers like Random Forest (RF), k-nearest neighbors (KNN), and Logistic Regression (LR). 11 In follow-up work, 15 used time series classi-ers like Convolutional Neural Networks (CNN) to obtain improved detection capabilities for explosives and their precursors using the kinetic or time series nature of the data.…”